{ "dataset": "SafeLegalAI Legal AI Regulation Tracker", "license": "CC BY 4.0", "url": "https://safelegalai.com/regulation", "generated": "2026-09-04T14:34:52.273Z", "count": 324, "documents": [ { "id": "id-mahkamah-agung-letter-112-2026-ai-document-services", "url": "https://safelegalai.com/regulation/other/id-mahkamah-agung-letter-112-2026-ai-document-services", "title": "Letter No. 112/WKMA.Y/HK2.1/IX/2026: draft judgment-writing guidance and templates for the new Criminal Codes", "body": "Supreme Court of Indonesia, Deputy Chief Justice for Judicial Affairs [Wakil Ketua Mahkamah Agung Bidang Yudisial]", "bodyShort": "Mahkamah Agung", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality" ], "effectiveDate": "2026-09-02T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-09-02T00:00:00.000Z", "url": "https://www.mahkamahagung.go.id/media/14915" } ], "summary": "A Supreme Court of Indonesia letter of 2 September 2026 circulating draft judgment-writing guidance and templates for consultation. Its annex contains what appears to be the Indonesian judiciary's first written restriction on AI: draft templates must not be uploaded to AI document-processing services.", "keyProvisions": [ "Annex section C.3 bars circulating the draft outside the working team, including uploading it to social media, open groups, public file-sharing pages, or document-processing and artificial intelligence services.", "Consultation period closes 18 September 2026." ], "relatedIncidents": [], "article": "indonesia-supreme-court-ai-judgment-drafting-guidance-2026-09", "sources": [ { "label": "Mahkamah Agung — letter and annex (official PDF)", "url": "https://www.mahkamahagung.go.id/media/14915" }, { "label": "Mahkamah Agung — covering letter", "url": "https://www.mahkamahagung.go.id/media/14914" } ], "country": "ID", "categories": [ "confidentiality-client-data", "judicial-use", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "in-sc-vijay-gadiya-v-union-of-india-2026", "url": "https://safelegalai.com/regulation/india/in-sc-vijay-gadiya-v-union-of-india-2026", "title": "Vijay Ghanshyam Gadiya v Union of India, 2026 INSC 947", "body": "Supreme Court of India", "bodyShort": "Supreme Court of India", "jurisdiction": "india", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff", "all" ], "requirements": [ "verification", "prohibition" ], "effectiveDate": "2026-09-02T00:00:00.000Z", "versions": [ { "version": "Judgment", "date": "2026-09-02T00:00:00.000Z", "url": "https://api.sci.gov.in/supremecourt/2026/21991/21991_2026_7_1501_73957_Judgement_02-Sep-2026.pdf", "changes": "Original judgment (Datta and Nagu JJ)." } ], "summary": "Supreme Court judgment of 2 September 2026 setting aside a customs penalty of Rs 425.27 crore because the Additional Commissioner's order relied on AI-generated citations that were non-existent or misrepresented. AI may serve as training wheels but entrusting it with the pilot's seat would be imprudent and dangerous; the matter was remanded to a different officer.", "keyProvisions": [ "Assistance can never be substituted for adjudication; AI cannot replace human decision-making by judicial or quasi-judicial authorities.", "Orders relying on fabricated or misrepresented AI citations are unsustainable.", "The appointing authority is to consider disciplinary action against the officer who authored the order." ], "relatedIncidents": [], "article": "india-supreme-court-gadiya-ai-citations-2026-09", "sources": [ { "label": "Supreme Court of India — Vijay Ghanshyam Gadiya v Union of India, 2026 INSC 947 (official PDF)", "url": "https://api.sci.gov.in/supremecourt/2026/21991/21991_2026_7_1501_73957_Judgement_02-Sep-2026.pdf" }, { "label": "Verdictum report and judgment PDF", "url": "https://www.verdictum.in/supreme-court/vijay-ghanshyam-gadiya-v-union-of-india-2026-insc-947-ai-customs-authority-penalty-1621205" }, { "label": "LiveLaw report", "url": "https://www.livelaw.in/supreme-court/supreme-court-sets-aside-425-crore-penalty-as-customs-authority-used-ai-generated-fake-case-laws-548329" } ], "country": "IN", "categories": [ "ai-decision-prohibition", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "lt-advokatura-ai-training-2026", "url": "https://safelegalai.com/regulation/other/lt-advokatura-ai-training-2026", "title": "AI in the advocate's practice: opportunities, tools and safe use [DI advokato praktikoje: galimybės, įrankiai ir saugus naudojimas]", "body": "Lithuanian Bar Association [Lietuvos advokatūra], Information Technology and Artificial Intelligence Committee", "bodyShort": "Lietuvos advokatūra", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "competence", "permissive" ], "versions": [ { "version": "1.0", "date": "2026-09-01T00:00:00.000Z", "url": "https://www.advokatura.lt/naujienos/rugsejo-9-diena-mokymai-tema-di-advokato-praktikoje-galimybes-irankiai-ir-saugus-naudojimas1/", "changes": "Announced 1 September 2026; full-day practical training held 9 September 2026 at the Advocates' House, Vilnius, organised by the Bar's IT and AI Committee with partners. Voluntary; advance registration and a personal laptop required." } ], "summary": "A voluntary full-day practical course run by the Lithuanian Bar's Information Technology and Artificial Intelligence Committee on 9 September 2026. It is the Bar's first substantive public AI initiative for advocates and stands in place of any published guidance document, which Lithuania still lacks.", "keyProvisions": [ "Covers AI tools, their practical application in daily advocate work, and the principles of safe, responsible and professional use.", "Aimed at applying skills in real practice: drafting documents, analysing information, organising daily processes and assessing AI's opportunities and associated risks.", "Organised by the Bar's standing Information Technology and Artificial Intelligence Committee, an advisory body to the Council of Advocates.", "Voluntary and fee-based via the Bar's shop; not a continuing-education mandate." ], "relatedIncidents": [], "sources": [ { "label": "Lietuvos advokatūra — training announcement (1 Sept 2026)", "url": "https://www.advokatura.lt/naujienos/rugsejo-9-diena-mokymai-tema-di-advokato-praktikoje-galimybes-irankiai-ir-saugus-naudojimas1/" } ], "country": "LT", "categories": [ "training-cle", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "jp-supreme-court-civil-trial-ai-pilot-2026", "url": "https://safelegalai.com/regulation/other/jp-supreme-court-civil-trial-ai-pilot-2026", "title": "Supreme Court of Japan generative-AI demonstration for civil trials (FY2027 budget request)", "body": "Supreme Court of Japan", "bodyShort": "Supreme Court of Japan", "jurisdiction": "other", "type": "policy", "status": "proposed", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality" ], "versions": [ { "version": "Budget request announcement", "date": "2026-08-28T00:00:00.000Z", "url": "https://news.web.nhk/newsweb/na/nd-20260828de46923", "changes": "Announced FY2027 budget request for system development and a demonstration using settled-case records." } ], "summary": "The Supreme Court of Japan announced on 28 August 2026 a budget request for the 2027 financial year to build and test a generative-AI system on records of settled civil cases. AI would draft record summaries and issue lists to reduce judges' workload, with leak prevention for personal data; AI is excluded from judges' decision-making.", "keyProvisions": [ "Demonstration limited to summarising records and organising issues, not decision-making.", "Security controls to prevent leakage of party data and trade secrets.", "Follows a January-February 2026 study by six judges using mock records, and full digitisation of civil procedure in May 2026." ], "relatedIncidents": [], "sources": [ { "label": "NHK (28 August 2026)", "url": "https://news.web.nhk/newsweb/na/nd-20260828de46923" }, { "label": "Kyodo via Yahoo Japan (22 March 2026)", "url": "https://news.yahoo.co.jp/articles/18ccb05e6341fdf68f83d05da47c66d500f73f42" }, { "label": "Japan Times (1 September 2026)", "url": "https://www.japantimes.co.jp/news/2026/09/01/japan/japan-supreme-court-ai-use/" } ], "country": "JP", "categories": [ "court-own-use", "judicial-use", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "tr-tbb-ai-guide-lawyers", "url": "https://safelegalai.com/regulation/other/tr-tbb-ai-guide-lawyers", "title": "Recommendation Guide on the Use of Artificial Intelligence for Lawyers [Avukatlar İçin Yapay Zekâ Kullanımı Tavsiye Rehberi]", "body": "Union of Turkish Bar Associations [Türkiye Barolar Birliği], Information and Technology Law Commission", "bodyShort": "Türkiye Barolar Birliği (TBB)", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms", "all" ], "requirements": [ "verification", "confidentiality", "consent", "competence", "supervision", "record-keeping", "prohibition", "disclosure" ], "versions": [ { "version": "First edition, 2026", "date": "2026-08-28T00:00:00.000Z", "url": "https://d.barobirlik.org.tr/2026/yapayzekatavsiyerehberi.pdf", "changes": "First national Turkish bar instrument on AI. Product of the TBB 'AI and Advocacy Workshop', Ankara, 14-15 February 2026. Foreword by TBB President R. Erinç Sağkan; presented by Commission spokesperson Dr Resul Göksoy." } ], "summary": "A 44-page principle-based, technology-neutral guide covering lawyers, trainee lawyers and law-office staff. It classifies AI use into risk tiers, sets duties of human oversight, verification, professional secrecy, data minimisation and client information, and adds practical annexes: vendor criteria, model office-policy clauses and client-engagement wording. Advisory, reviewed periodically by TBB.", "keyProvisions": [ "Final responsibility rests with the lawyer: AI output is never a final legal opinion and must not be given to a client, court, authority or third party unchecked.", "Risk-based classification by data type, impact on the client, proximity to the legal outcome, whether the tool is open or closed, cross-border transfer risk and contractual safeguards: free assistive use, use subject to safeguards and where needed consent, and prohibited uses.", "No legislation, precedent, citation or fact from an AI tool may be filed with a court, arbitrator, mediator or administrative body without confirmation against official sources; hallucinated citations may be a grave breach of professional care, and errors discovered later must be corrected promptly.", "Non-anonymised client data, case and enforcement file content, evidence, special-category data, trade secrets and negotiation strategy must not be entered into public or contractually unprotected AI systems.", "Clients must be informed, and where needed give consent, whenever AI use touches their data, strategy or the outcome of the service; a client may ask that AI not be used on their file.", "Time saved by AI must not be billed as lawyer time, and AI or general software costs must not be passed on without prior agreement; AI use is not itself a ground to reduce the fee.", "Profiling, scoring or predicting individual judges' decisions is incompatible with professional practice, as are social scoring, manipulative techniques and emotion recognition.", "Marketing claims such as guaranteed results, predicting court outcomes or lawyer-free legal advice are barred, as is facilitating non-lawyers or systems that appear to give independent legal advice.", "Vendor criteria table covering data provenance, data location and cross-border transfer, a contractual undertaking that prompts and documents are not used for model training, security controls, transparency on error rates and a written data-processor agreement.", "AI literacy modules to be built into trainee education and in-service training, with a TBB-coordinated trainer pool and materials standard." ], "relatedIncidents": [], "sources": [ { "label": "TBB Guide (official PDF)", "url": "https://d.barobirlik.org.tr/2026/yapayzekatavsiyerehberi.pdf" }, { "label": "TBB announcement (28 Aug 2026)", "url": "https://www.barobirlik.org.tr/Haberler/tbb-avukatlikta-yapay-zeka-kullanimina-iliskin-tavsiye-rehberini-yayimladi-86648" } ], "country": "TR", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "client-consent-transparency", "competence-duty", "supervision", "fees-billing", "unauthorised-practice", "training-cle", "procurement-vendor", "disclosure-filings", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "vn-decision-1671-national-ai-strategy", "url": "https://safelegalai.com/regulation/other/vn-decision-1671-national-ai-strategy", "title": "Decision No. 1671/QD-TTg approving the National Strategy on Artificial Intelligence to 2030, vision to 2045", "body": "Prime Minister of Vietnam [Thu tuong Chinh phu]", "bodyShort": "Prime Minister of Vietnam", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "permissive" ], "effectiveDate": "2026-08-28T00:00:00.000Z", "versions": [ { "version": "Signed", "date": "2026-08-28T00:00:00.000Z", "url": "https://vanban.chinhphu.vn/?pageid=27160&docid=219354", "changes": "Supersedes Decision No. 127/QD-TTg as Vietnam's national AI strategy." } ], "summary": "Vietnam's current national AI strategy, approved by the Prime Minister on 28 August 2026 and running to 2030 with a vision to 2045. It supersedes Decision No. 127/QD-TTg. It is a development strategy rather than a regulatory instrument and does not address artificial intelligence in legal practice.", "keyProvisions": [ "Approves the national AI strategy to 2030 with a vision to 2045.", "Supersedes Decision No. 127/QD-TTg.", "Development-oriented; contains no provisions on courts or legal practice." ], "relatedIncidents": [], "sources": [ { "label": "Government gazette portal — Decision 1671/QD-TTg", "url": "https://vanban.chinhphu.vn/?pageid=27160&docid=219354" } ], "country": "VN", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "sra-misuse-ai-warning-notice", "url": "https://safelegalai.com/regulation/uk/sra-misuse-ai-warning-notice", "title": "Misuse of AI: Warning notice", "body": "Solicitors Regulation Authority", "bodyShort": "SRA", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "supervision", "competence" ], "effectiveDate": "2026-08-17T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-08-17T00:00:00.000Z", "url": "https://www.sra.org.uk/solicitors/guidance/misuse-ai/" } ], "summary": "The SRA's first warning notice devoted to AI, issued 17 August 2026 after 42 reports of potential misuse in a year. It tells solicitors in England and Wales they remain accountable for AI output, must verify that cited authorities are genuine and relevant, must protect client data entered into AI tools, and must have governance and supervision controls.", "keyProvisions": [ "Solicitors are accountable for the accuracy of their work regardless of the tool used to produce it.", "Authorities relied on must be checked to be genuine, relevant and verifiably cited before use.", "Client information may only be entered into AI tools with contractual, technical and organisational safeguards, and not for model training without authority.", "Firms must have governance, systems and controls for AI use; COLPs must ensure supervisory arrangements cover AI-assisted work.", "Cross-refers to the SRA Effective Supervision guidance, Risk Outlook on AI and Balancing Duties in Litigation guidance." ], "relatedIncidents": [ "ayinde-v-haringey", "cork-v-smith", "bandla-v-sra" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "SRA — Misuse of AI warning notice (official page)", "url": "https://www.sra.org.uk/solicitors/guidance/misuse-ai/" }, { "label": "SRA press release — Responsible use of AI", "url": "https://www.sra.org.uk/news/news/releases/responsible-use-ai/" } ], "country": "GB", "categories": [ "verification-duty", "confidentiality-client-data", "supervision", "competence-duty", "bar-guidance", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ct-ai-legal-system-committee-final-report", "url": "https://safelegalai.com/regulation/us-state/us-ct-ai-legal-system-committee-final-report", "title": "Connecticut Judicial Branch, Committee on Artificial Intelligence in the Connecticut Legal System: Final Report", "body": "Connecticut Judicial Branch, Committee on Artificial Intelligence in the Connecticut Legal System", "bodyShort": "Connecticut Judicial Branch", "jurisdiction": "us-state", "region": "Connecticut", "type": "report", "status": "in-force", "appliesTo": [ "judges", "court-staff", "lawyers" ], "requirements": [ "verification", "confidentiality", "competence" ], "versions": [ { "version": "Final Report", "date": "2026-08-10T00:00:00.000Z", "url": "https://www.jud.ct.gov/committees/artificial_intelligence_CLS/FinalReport_R.pdf", "changes": "Final report of the committee" } ], "summary": "The committee established by the Connecticut Judicial Branch to examine AI in the legal system closed with a final report on 10 August 2026. It records that Practice Book section 4-9 and related amendments discharged the rules charge, sets eight guiding principles for judicial AI use, and describes the Branch's permanent AI governance structure.", "keyProvisions": [ "The committee concluded no further court rule or evidence code recommendations were needed after adoption of Practice Book section 4-9, amendments to section 4-2 and corresponding appellate rules amendments", "Guiding principles for Branch AI use: judicial independence, fairness and impartiality, transparency, accountability, protection of confidential information, cybersecurity, human oversight of AI-generated work product, responsible innovation", "The Judicial Branch established an Internal Artificial Intelligence Committee and a standing AI Review Subcommittee to review proposed AI software and hardware before implementation", "The Connecticut Code of Evidence Oversight Committee, chaired by Judge Hope C. Seeley, is reviewing whether the evidence rules need amendment for generative AI", "AI should augment, not replace, human judgment, and AI-derived information should be verified by traditional methods" ], "relatedIncidents": [], "sources": [ { "label": "Final Report (PDF)", "url": "https://www.jud.ct.gov/committees/artificial_intelligence_CLS/FinalReport_R.pdf" }, { "label": "Committee page", "url": "https://www.jud.ct.gov/committees/artificial_intelligence_CLS/default.htm" } ], "country": "US", "categories": [ "judicial-use", "court-own-use", "evidence-admissibility", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "id-peradi-pedoman-ai-advokat", "url": "https://safelegalai.com/regulation/other/id-peradi-pedoman-ai-advokat", "title": "Guideline on the Use of Artificial Intelligence for Indonesian Advocates [Pedoman Penggunaan Kecerdasan Buatan]", "body": "Indonesian Bar Association [Perhimpunan Advokat Indonesia, PERADI]", "bodyShort": "PERADI", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "supervision", "prohibition", "competence" ], "effectiveDate": "2026-08-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-08-06T00:00:00.000Z", "url": "https://peradi.id/peradi-terbitkan-pedoman-etika-ai-pertama-di-indonesia-perkokoh-integritas-advokat-di-era-digital/", "changes": "The guideline text itself is not published at an official URL; content is taken from PERADI's own announcement." } ], "summary": "Indonesia's first bar AI guideline, issued by PERADI on 6 August 2026. It uses permitted, supervised and prohibited risk tiers, treats AI as augmentation rather than a replacement for human reasoning, and requires every AI output to pass through the advocate's own intellectual review before use.", "keyProvisions": [ "Every AI output must undergo the advocate's own intellectual examination and filtration; professional accountability remains with the advocate.", "AI must be an assistant or augmentation, not a replacement for human reasoning.", "Prohibits fully automated legal decisions and any use that eliminates human control, misleads clients or courts, breaches confidentiality, or discriminates.", "Firms must continue to give trainee advocates adequate mentoring, training and learning opportunities; AI must not replace mentorship.", "PERADI commits to encouraging continuing education so that members have adequate technology literacy." ], "relatedIncidents": [], "sources": [ { "label": "PERADI — announcement of the guideline (6 Aug 2026)", "url": "https://peradi.id/peradi-terbitkan-pedoman-etika-ai-pertama-di-indonesia-perkokoh-integritas-advokat-di-era-digital/" } ], "country": "ID", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "ai-decision-prohibition", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ireland-high-court-pd-hc142-generative-ai", "url": "https://safelegalai.com/regulation/ireland/ireland-high-court-pd-hc142-generative-ai", "title": "Practice Direction HC 142: Responsible Use of Generative Artificial Intelligence in Court Documents", "body": "High Court of Ireland", "bodyShort": "Irish High Court", "jurisdiction": "ireland", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "certification", "disclosure", "prohibition" ], "effectiveDate": "2026-09-01T00:00:00.000Z", "versions": [ { "version": "Signed 29 July 2026, in operation 1 September 2026", "date": "2026-07-29T00:00:00.000Z", "url": "https://www.courts.ie/practice-directions/full-practice-direction?url=practice-direction-on-the-responsible-use-of-generative-artificial-intelligence-in-court-documents" } ], "summary": "Ireland's first court rule on AI, signed by the President of the High Court on 29 July 2026 and in operation from 1 September 2026. It covers all civil documents and requires independent human verification of every AI-assisted citation, a declaration in affidavits and witness statements that generative AI was not used for substantive content, and disclosure for expert reports.", "keyProvisions": [ "Applies to pleadings, submissions, affidavits, witness statements and expert reports prepared after 1 September 2026.", "Every AI-assisted citation or proposition must be independently verified by a human; checking with another AI tool is insufficient.", "Affidavits and witness statements must carry a declaration that generative AI was not used to generate substantive content.", "Expert reports require a no-substantive-use declaration or disclosure of AI use.", "Sanctions include rejection of documents, adverse or personal costs, strike-out and referral to the regulator." ], "relatedIncidents": [ "guerin-v-odoherty", "von-geitz-v-kelly" ], "article": "ireland-ai-guidance-history", "sources": [ { "label": "Courts Service of Ireland — Practice Direction HC 142", "url": "https://www.courts.ie/practice-directions/full-practice-direction?url=practice-direction-on-the-responsible-use-of-generative-artificial-intelligence-in-court-documents" } ], "country": "IE", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person", "sanctions-record", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eu-ai-act", "url": "https://safelegalai.com/regulation/eu/eu-ai-act", "title": "Regulation (EU) 2024/1689 (Artificial Intelligence Act)", "body": "European Parliament and Council of the European Union", "bodyShort": "EU", "jurisdiction": "eu", "type": "statute", "status": "in-force", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "risk-classification", "prohibition", "disclosure", "competence" ], "effectiveDate": "2024-08-01T00:00:00.000Z", "versions": [ { "version": "As amended by Regulation (EU) 2026/1744", "date": "2026-07-27T00:00:00.000Z", "url": "https://eur-lex.europa.eu/eli/reg/2024/1689/oj", "changes": "Digital Omnibus on AI defers Annex III high-risk obligations to 2 December 2027 and Annex I product-embedded obligations to 2 August 2028; Article 50 transparency and GPAI obligations unchanged." }, { "version": "Original (OJ L, 12 July 2024)", "date": "2024-07-12T00:00:00.000Z", "url": "https://eur-lex.europa.eu/eli/reg/2024/1689/oj" } ], "summary": "The EU AI Act sets risk-based obligations for providers and deployers of AI systems, including law firms and legal tech vendors operating in the European Union. Prohibited practices and AI literacy duties applied from 2 February 2025, GPAI rules from 2 August 2025 and Article 50 transparency from 2 August 2026, with high-risk obligations now deferred to 2027 and 2028.", "keyProvisions": [ "Article 4 requires providers and deployers to ensure sufficient AI literacy among staff, applicable from 2 February 2025.", "Article 5 prohibited practices applied from 2 February 2025.", "General-purpose AI model obligations applied from 2 August 2025; Commission enforcement powers from 2 August 2026.", "Article 50 transparency obligations, including marking of synthetic content, applied from 2 August 2026.", "Annex III stand-alone high-risk obligations apply from 2 December 2027; Annex I product-embedded systems from 2 August 2028." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "EUR-Lex — Regulation (EU) 2024/1689", "url": "https://eur-lex.europa.eu/eli/reg/2024/1689/oj" }, { "label": "EUR-Lex — Regulation (EU) 2026/1744 (amending act)", "url": "https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng" } ], "country": "EU", "categories": [ "ai-act-horizontal", "judicial-use", "data-protection", "evidence-admissibility", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "eu-digital-omnibus-ai-2026-1744", "url": "https://safelegalai.com/regulation/eu/eu-digital-omnibus-ai-2026-1744", "title": "Regulation (EU) 2026/1744 (Digital Omnibus on AI)", "body": "European Parliament and Council of the European Union", "bodyShort": "EU", "jurisdiction": "eu", "type": "statute", "status": "in-force", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2026-07-27T00:00:00.000Z", "versions": [ { "version": "Official Journal text", "date": "2026-07-24T00:00:00.000Z", "url": "https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng" } ], "summary": "Regulation (EU) 2026/1744, adopted 8 July 2026 and in force from 27 July 2026, amends the EU AI Act, the EASA Regulation and the Machinery Regulation. It defers Annex III high-risk obligations to 2 December 2027 and Annex I obligations to 2 August 2028. Article 50 transparency and GPAI duties were not deferred and applied from 2 August 2026.", "keyProvisions": [ "Annex III stand-alone high-risk AI obligations deferred to 2 December 2027.", "Annex I product-embedded high-risk obligations deferred to 2 August 2028.", "Commission must adopt delegated acts by 2 August 2027 and post-market-monitoring guidance by 2 September 2027.", "Article 50 transparency obligations and GPAI obligations were not deferred.", "Published in the Official Journal on 24 July 2026." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "EUR-Lex — Regulation (EU) 2026/1744", "url": "https://eur-lex.europa.eu/eli/reg/2026/1744/oj/eng" } ], "country": "EU", "categories": [ "ai-act-horizontal", "judicial-use" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "eu-commission-article-50-transparency-guidelines", "url": "https://safelegalai.com/regulation/eu/eu-commission-article-50-transparency-guidelines", "title": "Guidelines on the transparency obligations under Article 50 of the AI Act", "body": "European Commission", "bodyShort": "European Commission", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "disclosure" ], "effectiveDate": "2026-07-20T00:00:00.000Z", "versions": [ { "version": "Final, C(2026) 5054 final", "date": "2026-07-20T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/library/guidelines-transparency-obligations-providers-and-deployers-ai-systems", "changes": "Final version, approved by the Commission on 20 July 2026 (C(2026) 5054 final), replaces the 8 May 2026 draft." }, { "version": "Draft for consultation", "date": "2026-05-08T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content" } ], "summary": "European Commission guidelines, approved 20 July 2026 as C(2026) 5054 final, on the EU AI Act Article 50 transparency duties that applied from 2 August 2026. Providers mark synthetic output in a machine-readable, detectable way; deployers disclose deepfakes and AI-generated text on matters of public interest, unless the text has had substantive human review.", "keyProvisions": [ "Providers must mark AI-generated output so it is machine-readable and detectable.", "Deployers must disclose deepfakes and AI-generated text published on matters of public interest.", "Obligations apply from 2 August 2026; under new Article 111(4) of the AI Act, generative systems placed on the market before that date must meet Article 50(2) marking by 2 December 2026.", "Paragraphs 134 to 136 define the human-review exception for public-interest text: substantive examination by a person with relevant knowledge, fact-checking as a minimum, and no AI edits after sign-off.", "Guidelines are interpretive and accompany the Code of Practice on AI-generated content." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — Guidelines on transparency obligations for providers and deployers of AI systems (library page, 20 July 2026)", "url": "https://digital-strategy.ec.europa.eu/en/library/guidelines-transparency-obligations-providers-and-deployers-ai-systems" }, { "label": "European Commission — Guidelines on transparency obligations policy page", "url": "https://digital-strategy.ec.europa.eu/en/policies/guidelines-transparency-ai-generated-content" } ], "country": "EU", "categories": [ "ai-act-horizontal", "client-consent-transparency", "evidence-admissibility" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "gr-law-5321-2026-ai-act-implementation", "url": "https://safelegalai.com/regulation/other/gr-law-5321-2026-ai-act-implementation", "title": "Law 5321/2026: measures for the implementation of Regulation (EU) 2024/1689 and amendment of Law 4961/2022", "body": "Hellenic Parliament (Vouli ton Ellinon)", "bodyShort": "Parliament (GR)", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Government Gazette 2026 (FEK id 802524)", "date": "2026-07-20T00:00:00.000Z", "url": "https://search.et.gr/el/fek/?fekId=802524", "changes": "Original. Published in Government Gazette A 114 of 20 July 2026 (FEK A 114/20.07.2026); date confirmed from the Special Secretariat for AI announcement and Greek legal databases on 4 September 2026." } ], "summary": "Greek implementing act for the EU AI Act, listed by the Special Secretariat for AI and Data Governance. It designates the Hellenic Data Protection Authority as principal market-surveillance authority and EETT as notifying authority, and amends the 2022 emerging-technologies law that already required public-sector algorithmic impact assessments and AI registers. No justice-specific provisions were identified.", "keyProvisions": [ "Designates national competent authorities for the EU AI Act, with the HDPA leading market surveillance.", "Names EETT as notifying authority.", "Amends Law 4961/2022 on emerging technologies, which requires public-sector algorithmic impact assessments and AI registers." ], "relatedIncidents": [], "sources": [ { "label": "Special Secretariat for AI and Data Governance — resources", "url": "https://ai.gov.gr/en/resources/" }, { "label": "Special Secretariat for AI: Greece adopts national framework for the AI Act (announcement)", "url": "https://ai.gov.gr/announcements/i-ellada-apokta-to-ethniko-plaisio-efarmogis-toy-ai-act/" }, { "label": "Government Gazette record (FEK id 802524)", "url": "https://search.et.gr/el/fek/?fekId=802524" }, { "label": "Legal 500 Greece AI chapter (secondary)", "url": "https://www.legal500.com/guides/chapter/greece-artificial-intelligence/" } ], "country": "GR", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "in-bci-digital-conduct-circular-2026", "url": "https://safelegalai.com/regulation/india/in-bci-digital-conduct-circular-2026", "title": "Bar Council of India circular on digital and social-media conduct and professional responsibility of advocates", "body": "Bar Council of India", "bodyShort": "BCI", "jurisdiction": "india", "type": "regulation", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "disclosure", "prohibition" ], "effectiveDate": "2026-07-17T00:00:00.000Z", "versions": [ { "version": "Original (reported as BCI/D/4659/2026)", "date": "2026-07-17T00:00:00.000Z", "url": "https://www.verdictum.in/news/bar-council-of-india-digital-social-media-conduct-legal-practitioners-1617984", "changes": "Original circular; official PDF not located on barcouncilofindia.org." } ], "summary": "Bar Council of India circular of 17 July 2026 on advocates', law students' and interns' online conduct. Advocates publishing legal-awareness content must add disclaimers and accurately disclose any use of generative-AI tools; fabricated judgments, fake citations, AI-generated non-existent case law and deepfake courtroom content are prohibited, with disciplinary referral or contempt where warranted.", "keyProvisions": [ "Accurate disclosure of generative-AI use in online legal content.", "Prohibition on fabricated judgments, fake citations, unverified screenshots and AI-generated non-existent case law.", "Prohibition on deepfakes, voice cloning and synthetic content involving judges, lawyers or litigants.", "Violations may lead to disciplinary action, referral to State Bar Councils or contempt proceedings." ], "relatedIncidents": [], "sources": [ { "label": "Verdictum report", "url": "https://www.verdictum.in/news/bar-council-of-india-digital-social-media-conduct-legal-practitioners-1617984" } ], "country": "IN", "categories": [ "bar-guidance", "evidence-admissibility", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "za-judiciary-conference-2026-resolutions-ai", "url": "https://safelegalai.com/regulation/south-africa/za-judiciary-conference-2026-resolutions-ai", "title": "2026 Judiciary Conference Resolutions, Section 5 — Information Communication Technology, Digitisation and Artificial Intelligence", "body": "South African Judiciary (2026 Judiciary Conference, Umhlanga, Durban)", "bodyShort": "SA Judiciary Conference", "jurisdiction": "south-africa", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "competence", "record-keeping", "verification" ], "versions": [ { "version": "Final adopted", "date": "2026-07-16T00:00:00.000Z", "url": "https://www.judiciary.org.za/images/news/2026/FINAL%20ADOPTED%202026%20Judiciary%20Conference%20Resolutions.pdf", "changes": "Adopted at the 2026 Judiciary Conference held in Umhlanga, Durban, 13–16 July 2026." } ], "summary": "Resolutions adopted by the South African Judiciary on 16 July 2026. Section 5 directs the Judiciary Administration and Information Technology Steering Committee to finalise the draft judiciary AI policy, requires the Judiciary to retain governance over AI procurement and deployment, restricts initial AI use to efficiency tools, and mandates ongoing judicial education on responsible AI use.", "keyProvisions": [ "5.2 — the Steering Committee is to finalise the Draft Policy on Artificial Intelligence for adoption, defining with greater specificity the permissible and prohibited uses of AI and the distinction between administrative and adjudicative functions, to safeguard confidentiality and the integrity of the administration of justice.", "5.3 — the Judiciary is to retain ownership and governance over the development, procurement and use of AI systems, keeping human oversight, judicial reasoning and judicial determination central, with training data reflecting South Africa's constitutional values.", "5.4 — AI should be leveraged through customised and secure platforms, initially for court recording and transcription, translation, case management and document summarisation.", "5.5 — the Judiciary should provide ongoing education on responsible AI use, including prompt design and verification, while maintaining legal research, analytical, reasoning and judgment-writing skills.", "5.1 — modernisation of court infrastructure, integrated case management and electronic filing, supported by cybersecurity and ICT capacity." ], "relatedIncidents": [], "sources": [ { "label": "South African Judiciary — 2026 Judiciary Conference Resolutions (official PDF)", "url": "https://www.judiciary.org.za/images/news/2026/FINAL%20ADOPTED%202026%20Judiciary%20Conference%20Resolutions.pdf" } ], "country": "ZA", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "training-cle", "procurement-vendor", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ec-cofj-reform-bill-ai-2026", "url": "https://safelegalai.com/regulation/other/ec-cofj-reform-bill-ai-2026", "title": "Bill reforming the Organic Code of the Judicial Function to regulate AI use in judicial proceedings", "body": "National Assembly of Ecuador (Asamblea Nacional)", "bodyShort": "Asamblea Nacional", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "judges", "court-staff", "lawyers" ], "requirements": [ "verification" ], "versions": [ { "version": "Presented by Asambleista Monica Salazar", "date": "2026-07-13T00:00:00.000Z", "url": "https://www.infobae.com/america/america-latina/2026/07/13/ecuador-busca-regular-el-uso-de-inteligencia-artificial-en-la-justicia-con-un-proyecto-que-fija-limites-y-sanciones/", "changes": "Introduced; awaiting Legislative Administration Council review. No official URL located." } ], "summary": "Bill amending the Organic Code of the Judicial Function so that judges and judicial officials may not use AI-generated information without first verifying its accuracy, relevance and currency; lawyers may not file AI-generated content without verifying it where it could mislead; and incorporating unverified AI content becomes a serious disciplinary offence for judicial personnel.", "keyProvisions": [ "Verification duty for judges before using AI-generated information.", "Verification duty for lawyers filing AI-generated content.", "Serious disciplinary offence for judicial personnel who incorporate unverified AI output.", "Judges remain responsible for signed documents regardless of AI assistance." ], "relatedIncidents": [], "sources": [ { "label": "Infobae report", "url": "https://www.infobae.com/america/america-latina/2026/07/13/ecuador-busca-regular-el-uso-de-inteligencia-artificial-en-la-justicia-con-un-proyecto-que-fija-limites-y-sanciones/" } ], "country": "EC", "categories": [ "verification-duty", "sanctions-record", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "int-uncitral-wgii-ai-dispute-resolution-colloquium", "url": "https://safelegalai.com/regulation/international/int-uncitral-wgii-ai-dispute-resolution-colloquium", "title": "UNCITRAL Working Group II exploratory work on the use of AI in dispute resolution", "body": "United Nations Commission on International Trade Law, Working Group II (Dispute Settlement)", "bodyShort": "UNCITRAL WG II", "jurisdiction": "international", "type": "consultation", "status": "consultation", "appliesTo": [ "parties", "lawyers" ], "requirements": [ "permissive" ], "versions": [ { "version": "Commission 59th session", "date": "2026-07-10T00:00:00.000Z", "url": "https://unis.unvienna.org/unis/en/pressrels/2026/unisl400.html", "changes": "Commission adopted texts on electronic arbitral awards and Model Law revisions and set future work on dispute resolution in the digital economy; no AI-specific text adopted." }, { "version": "Colloquium on AI in dispute resolution and remote hearings", "date": "2026-02-16T00:00:00.000Z", "url": "https://uncitral.un.org/en/colloquium_aidisputeresolution" } ], "summary": "UNCITRAL has no AI-in-arbitration instrument. On a 2025 Commission mandate, Working Group II held a colloquium on 16-17 February 2026 covering arbitrators' use of AI, AI in mediation and remote hearings, to inform a recommendation to the Commission's 59th session, which instead prioritised dispute resolution in the digital economy for future work.", "keyProvisions": [ "Panels covered arbitrators' use of AI and its impact on recognition and enforcement of awards.", "Further panels addressed AI in mediation and the conduct of remote hearings.", "Outputs are exploratory; any guidance text would come through later WG II sessions." ], "relatedIncidents": [], "sources": [ { "label": "UNCITRAL — colloquium page", "url": "https://uncitral.un.org/en/colloquium_aidisputeresolution" }, { "label": "UNIS press release on the 59th session", "url": "https://unis.unvienna.org/unis/en/pressrels/2026/unisl400.html" } ], "country": "INT", "categories": [ "arbitration-adr" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "my-naio-ai-governance-bill-consultation", "url": "https://safelegalai.com/regulation/other/my-naio-ai-governance-bill-consultation", "title": "Public Consultation: Proposed Artificial Intelligence (AI) Governance Bill", "body": "National AI Office (NAIO), Ministry of Digital", "bodyShort": "NAIO", "jurisdiction": "other", "type": "consultation", "status": "consultation", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2026-07-10T00:00:00.000Z", "versions": [ { "version": "Consultation open 10 July to 1 August 2026", "date": "2026-07-10T00:00:00.000Z", "url": "https://upc.mpc.gov.my/view-consultation/264" } ], "summary": "A pre-drafting public consultation run through Malaysia's Unified Public Consultation portal on a proposed AI Governance Bill, open from 10 July to 1 August 2026. No bill had been introduced in the Malaysian Parliament as at 4 September 2026, and justice or legal-sector uses are not separately addressed.", "keyProvisions": [ "Pre-drafting consultation on a proposed statutory AI governance framework.", "Consultation closed 1 August 2026; no bill introduced as at 4 September 2026." ], "relatedIncidents": [], "sources": [ { "label": "Unified Public Consultation — proposed AI Governance Bill", "url": "https://upc.mpc.gov.my/view-consultation/264" } ], "country": "MY", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eu-code-of-practice-ai-generated-content", "url": "https://safelegalai.com/regulation/eu/eu-code-of-practice-ai-generated-content", "title": "Code of Practice on Transparency of AI-generated Content", "body": "European AI Office (independent chairs)", "bodyShort": "EU AI Office", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "disclosure" ], "effectiveDate": "2026-06-10T00:00:00.000Z", "versions": [ { "version": "Commission adequacy opinion and AI Board endorsement", "date": "2026-07-09T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content", "changes": "Commission opinion (8 July 2026) and AI Board endorsement (9 July 2026) confirm the code as an adequate voluntary tool for Article 50(2), (4) and (5) compliance." }, { "version": "Final code text", "date": "2026-06-10T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content" } ], "summary": "A voluntary EU code, finalised 10 June 2026 and confirmed adequate by the Commission in July 2026, that shows providers and deployers how to meet the AI Act's Article 50 marking and labelling duties for AI-generated text, images, audio and video. Around 190 signatories had joined by late July 2026, with watermark-detection interoperability due by 2 February 2027.", "keyProvisions": [ "Covers Article 50(2) machine-readable marking of synthetic output by providers.", "Covers Article 50(4) deepfake disclosure and AI-generated public-interest text by deployers.", "Signatories commit to interoperable watermark detection by 2 February 2027.", "Confirmed as an adequate voluntary compliance tool by the Commission on 8 July 2026 and the AI Board on 9 July 2026." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — Code of Practice on AI-generated content", "url": "https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content" } ], "country": "EU", "categories": [ "ai-act-horizontal", "evidence-admissibility" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ca-bc-provincial-court-ai-guidance", "url": "https://safelegalai.com/regulation/canada/ca-bc-provincial-court-ai-guidance", "title": "Guidance on using AI to prepare for court", "body": "Provincial Court of British Columbia", "bodyShort": "BC Provincial Court", "jurisdiction": "canada", "region": "British Columbia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure", "confidentiality" ], "versions": [ { "version": "eNews guidance", "date": "2026-07-08T00:00:00.000Z", "url": "https://provincialcourt.bc.ca/news-notices-policies-and-practice-directions/enews/07-07-2026" } ], "summary": "Plain-language guidance published by the Provincial Court of British Columbia on 8 July 2026 for lawyers and self-represented litigants. It warns of hallucinated citations, requires independent verification, asks users to hyperlink authorities, requires an honest answer if a judge asks whether AI was used, and cautions against uploading personal information to AI tools.", "keyProvisions": [ "Independently verify any AI-produced legal references, statutes and judgments before relying on them.", "Hyperlink cited laws and decisions with section or paragraph references so the judge can check them.", "If a judge asks whether AI assisted in preparing the case, answer honestly; this applies to lawyers and self-represented litigants alike.", "Remove personal details before uploading documents to AI tools, which may retain them.", "Recognises accessibility benefits of AI tools for people with disabilities." ], "relatedIncidents": [], "sources": [ { "label": "Provincial Court of BC — Guidance on using AI to prepare for court", "url": "https://provincialcourt.bc.ca/news-notices-policies-and-practice-directions/enews/07-07-2026" } ], "country": "CA", "categories": [ "litigant-in-person", "verification-duty", "disclosure-filings", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "in-sc-pooja-ramesh-singh-v-jk-bank-2026", "url": "https://safelegalai.com/regulation/india/in-sc-pooja-ramesh-singh-v-jk-bank-2026", "title": "Pooja Ramesh Singh v Jammu and Kashmir Bank Ltd, 2026 INSC 668", "body": "Supreme Court of India", "bodyShort": "Supreme Court of India", "jurisdiction": "india", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "all" ], "requirements": [ "verification", "prohibition" ], "effectiveDate": "2026-07-02T00:00:00.000Z", "versions": [ { "version": "Judgment", "date": "2026-07-02T00:00:00.000Z", "url": "https://indiankanoon.org/doc/113338666/", "changes": "Original judgment (Narasimha and Aradhe JJ)." } ], "summary": "Supreme Court judgment of 2 July 2026 setting aside NCLT and NCLAT orders that relied on six non-existent, AI-hallucinated precedents. A decision resting on fake or hallucinated material is no decision in law; Bar and Bench owe a zero-tolerance duty to verify AI-generated authority, and the Bar Council of India was directed to frame guidelines and disciplinary measures.", "keyProvisions": [ "Orders founded even partly on fake or hallucinated precedents must be set aside.", "Advocates commit professional misconduct by citing AI-generated judgments without verification.", "Courts and tribunals must retain absolute control over AI in adjudication and must not trust authorities they cannot verify.", "Bar Council of India to constitute a committee on AI use by advocates and prescribe disciplinary consequences." ], "relatedIncidents": [], "sources": [ { "label": "Indian Kanoon text", "url": "https://indiankanoon.org/doc/113338666/" }, { "label": "iPleaders analysis of Indian AI-citation cases", "url": "https://blog.ipleaders.in/ai-hallucinated-case-law-fake-citations-india/" } ], "country": "IN", "categories": [ "verification-duty", "sanctions-record", "ai-decision-prohibition", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "no-hoyesterett-advokatveiledningen-ai", "url": "https://safelegalai.com/regulation/other/no-hoyesterett-advokatveiledningen-ai", "title": "Veiledning for advokater, section 19 'Bruk av KI-verktøy'", "body": "Supreme Court of Norway [Norges Høyesterett]", "bodyShort": "Norwegian Supreme Court", "jurisdiction": "other", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "barristers" ], "requirements": [ "verification", "competence" ], "versions": [ { "version": "Veiledning for advokater, per juli 2026", "date": "2026-07-01T00:00:00.000Z", "url": "https://www.domstol.no/contentassets/d1d57fd8703546fd8f618c124e32d1d3/advokatveiledningen-juli-2026.pdf", "changes": "Current published version; the guidance is updated several times a year. Month-only date: the source gives the month, so the 1st is recorded as the date." }, { "version": "AI section added after the April 2025 incident", "date": "2025-04-01T00:00:00.000Z", "url": "https://www.domstol.no/no/hoyesterett/advokater/advokatveiledningen/", "changes": "Section on AI tools introduced after the Court received a brief containing AI-fabricated statutes and preparatory works. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "The Supreme Court of Norway's guidance for lawyers appearing before it, which since April 2025 contains a dedicated section 19 on the use of AI tools. It was added after the Court received its first brief citing statutes and preparatory works that did not exist, and is cross-referenced from every pleading-drafting section of the guidance.", "keyProvisions": [ "A lawyer's duties are the same when using AI tools as otherwise, and the lawyer is personally responsible for all information used, including legal material.", "'Det er kjent at KI kan hallusinere og finne opp blant annet rettskilder. Kilder og andre opplysninger KI oppgir, må kvalitetssikres.'", "AI should only be used in areas where the lawyer has sufficient competence to check the content.", "The lawyer must satisfy themselves in the ordinary way that central sources have not been overlooked, and avoid unnecessary citations." ], "relatedIncidents": [], "sources": [ { "label": "Norges Høyesterett — Veiledning for advokater, July 2026 (PDF)", "url": "https://www.domstol.no/contentassets/d1d57fd8703546fd8f618c124e32d1d3/advokatveiledningen-juli-2026.pdf" }, { "label": "Norges Høyesterett — Advokatveiledningen", "url": "https://www.domstol.no/no/hoyesterett/advokater/advokatveiledningen/" } ], "country": "NO", "categories": [ "verification-duty", "competence-duty", "bar-guidance", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "za-lssa-revised-generative-ai-ethics-guidelines", "url": "https://safelegalai.com/regulation/south-africa/za-lssa-revised-generative-ai-ethics-guidelines", "title": "Revised Ethics Guidelines for Legal Practitioners in South Africa on the Use of Generative AI", "body": "Law Society of South Africa AI Committee (drafters, published via De Rebus)", "bodyShort": "LSSA AI Committee", "jurisdiction": "south-africa", "type": "guidance", "status": "proposed", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "competence", "confidentiality" ], "versions": [ { "version": "Revised draft", "date": "2026-07-01T00:00:00.000Z", "url": "https://www.derebus.org.za/responsible-ai-use-in-south-african-legal-practice-our-revised-guidelines-and-the-consultation-that-shaped-them/" } ], "summary": "Revised draft ethics guidelines on generative AI for South African legal practitioners, published 1 July 2026 in De Rebus by drafters working with the LSSA AI Committee after a 2026 consultation. Practitioners are fully responsible for AI-assisted work, must assess local relevance and bias, and must address cybersecurity risks. Not formally adopted by the LSSA or Legal Practice Council.", "keyProvisions": [ "Legal practitioners are fully responsible for AI-assisted work product.", "Output must be evaluated for accuracy, local legal relevance and bias.", "Practitioners must address cybersecurity risks such as prompt injection and data corruption.", "Offered as a collaborative contribution; not formally adopted by the LSSA or LPC." ], "relatedIncidents": [ "mavundla-v-mec-kzn", "northbound-processing", "parker-v-forsyth" ], "sources": [ { "label": "De Rebus — revised guidelines and consultation summary", "url": "https://www.derebus.org.za/responsible-ai-use-in-south-african-legal-practice-our-revised-guidelines-and-the-consultation-that-shaped-them/" } ], "country": "ZA", "categories": [ "bar-guidance", "verification-duty", "competence-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "br-tjpr-9-camara-ai-citation-fine-2026", "url": "https://safelegalai.com/regulation/brazil/br-tjpr-9-camara-ai-citation-fine-2026", "title": "TJPR 9th Civil Chamber: 2% bad-faith fine and OAB referral for AI-fabricated STJ precedent", "body": "Court of Justice of Parana (Tribunal de Justica do Parana)", "bodyShort": "TJPR", "jurisdiction": "brazil", "region": "Parana", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification" ], "effectiveDate": "2026-06-30T00:00:00.000Z", "versions": [ { "version": "Judgment (proc. 0108267-74.2025.8.16.0000)", "date": "2026-06-30T00:00:00.000Z", "url": "https://www.tjpr.jus.br/noticias/-/asset_publisher/9jZB/content/tjpr-multa-advogado-por-litig%C3%A2ncia-de-m%C3%A1-f%C3%A9-pelo-uso-de-ia-sem-zelo-e-cautela-/18319", "changes": "Initial decision; reported by the court and by Migalhas." } ], "summary": "Representative of Brazil's sanctions line: the Parana Court of Justice fined a lawyer 2 per cent of the updated case value under article 81 of the Code of Civil Procedure for citing a non-existent STJ ruling fabricated by AI, and referred him to the Parana Bar, holding that AI support does not relieve the duty of rigorous verification.", "keyProvisions": [ "Fine of 2 per cent of case value for bad-faith litigation (CPC arts 80-81).", "Referral to OAB/PR for disciplinary investigation.", "Use of AI does not remove the lawyer's verification duty.", "Consistent with TJSC (Feb 2025), the Londrina federal court (Jul 2025) and TST (Mar 2026) decisions." ], "relatedIncidents": [], "sources": [ { "label": "TJPR press release", "url": "https://www.tjpr.jus.br/noticias/-/asset_publisher/9jZB/content/tjpr-multa-advogado-por-litig%C3%A2ncia-de-m%C3%A1-f%C3%A9-pelo-uso-de-ia-sem-zelo-e-cautela-/18319" }, { "label": "Migalhas report", "url": "https://www.migalhas.com.br/quentes/459252/tj-pr-ve-ma-fe-e-multa-advogado-por-julgado-confeccionado-por-ia" } ], "country": "BR", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "uk-cjc-ai-court-documents-interim-findings", "url": "https://safelegalai.com/regulation/uk/uk-cjc-ai-court-documents-interim-findings", "title": "The Use of AI for the Preparation of Court Documents: Update on Consultation Findings", "body": "Civil Justice Council", "bodyShort": "CJC", "jurisdiction": "uk", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "permissive", "disclosure" ], "effectiveDate": "2026-06-30T00:00:00.000Z", "versions": [ { "version": "Interim update", "date": "2026-06-30T00:00:00.000Z", "url": "https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/latest-news/" } ], "summary": "Interim report from the Civil Justice Council for England and Wales, published 30 June 2026, summarising responses to its AI and court documents consultation. It records a strong consensus that no AI-specific formal requirements are needed for professional drafting, considers proportionate transparency for expert evidence, and flags concerns about AI-shaped witness statements and litigants in person.", "keyProvisions": [ "No new AI-specific rules recommended for pleadings and skeleton arguments prepared by lawyers.", "Proportionate transparency for AI use in expert evidence remains under consideration.", "Concern recorded about witness statements shaped by AI rather than in the witness's own words.", "A distinct approach for litigants in person is anticipated in the final report, expected later in 2026." ], "relatedIncidents": [], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "Civil Justice Council — latest news", "url": "https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/latest-news/" } ], "country": "GB", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "np-judiciary-ict-policy-2083", "url": "https://safelegalai.com/regulation/other/np-judiciary-ict-policy-2083", "title": "Judiciary Information and Communication Technology Policy, 2083 (2026)", "body": "Supreme Court of Nepal (full bench)", "bodyShort": "Supreme Court of Nepal", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "permissive", "competence" ], "effectiveDate": "2026-06-24T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-06-24T00:00:00.000Z", "url": "https://supremecourt.gov.np/court/public/media/2026_07/6be1c3d55afda34eb3b18eebd6bb0d08.pdf", "changes": "Approved by the Supreme Court full bench; document dated 2083-03-10 BS; adoption reported 3 July 2026." } ], "summary": "Nepal's judiciary-wide ICT policy approved by the Supreme Court full bench in mid-2026. Strategy 5.2 adopts AI as an innovation tool: AI-based systems will support drafting, research and case management, but AI shall not be used in a way that directly intrudes on judicial decisions, and third-party AI must be assessed for reliability and bias before use.", "keyProvisions": [ "5.2.1: AI-based systems for drafting and research support; no AI use that directly affects judicial decisions.", "5.2.2: internal or third-party AI systems assessed for reliability, transparency, non-bias and algorithmic processes.", "5.2.3: AI-assisted data entry in the case-management system.", "Awareness and training for judges, officials, Bar members and court users." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Nepal (official PDF, Nepali)", "url": "https://supremecourt.gov.np/court/public/media/2026_07/6be1c3d55afda34eb3b18eebd6bb0d08.pdf" }, { "label": "Kathmandu Post report", "url": "https://kathmandupost.com/national/2026/07/03/nepal-s-courts-embrace-ai-keep-judgment-writing-human" } ], "country": "NP", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ct-practice-book-4-9-genai-compliance", "url": "https://safelegalai.com/regulation/us-state/ct-practice-book-4-9-genai-compliance", "title": "Connecticut Practice Book Section 4-9: Generative Artificial Intelligence Compliance", "body": "Connecticut Judicial Branch (Rules Committee of the Superior Court)", "bodyShort": "Connecticut Judicial Branch", "jurisdiction": "us-state", "region": "Connecticut", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "verification", "certification" ], "effectiveDate": "2026-06-23T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-06-23T00:00:00.000Z", "url": "https://www.jud.ct.gov/legalresources/NewsLog/Post?id=6805" } ], "summary": "A statewide Connecticut court rule effective 23 June 2026 on publication in the Connecticut Law Journal. Anyone who uses generative AI to create or edit a filed document must independently verify every citation, authority and item of evidence, and the filer's signature certifies diligent compliance. Sanctions include nonsuit or default. It is among the first statewide binding verification rules.", "keyProvisions": [ "Applies to any filed document created or edited with generative AI, by lawyers or self-represented parties.", "Every citation, legal authority and item of evidence must be independently verified before filing.", "Filing constitutes a certification of good-faith, diligent compliance with the rule.", "Non-compliance may be sanctioned, including by nonsuit or default judgment." ], "relatedIncidents": [], "sources": [ { "label": "Connecticut Judicial Branch — news post on Practice Book Section 4-9", "url": "https://www.jud.ct.gov/legalresources/NewsLog/Post?id=6805" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-ri-supreme-court-order-genai-guidelines", "url": "https://safelegalai.com/regulation/us-state/us-ri-supreme-court-order-genai-guidelines", "title": "In re Amendments to Article V, Rule 1.1 of the Supreme Court Rules (Professional Conduct) and Adoption of Interim Generative AI Guidelines", "body": "Supreme Court of Rhode Island", "bodyShort": "RI Supreme Court", "jurisdiction": "us-state", "region": "Rhode Island", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "supervision", "consent" ], "versions": [ { "version": "Original", "date": "2026-06-17T00:00:00.000Z", "url": "https://www.courts.ri.gov/Miscellaneous%20Orders/Supreme-2026-06_In_reAmendments_Article_V_Rule_1_1_Professional_Conduct.pdf", "changes": "Amends Rule 1.1 Comment [6] and adopts interim generative AI guidelines for lawyers and for judicial officers" } ], "summary": "By order of 17 June 2026 the Rhode Island Supreme Court amended Article V, Rule 1.1 Comment [6] so that competence expressly includes keeping abreast of the benefits and risks of developing technology, and adopted advisory interim generative-AI guidelines for Rhode Island lawyers and for judicial officers. The guidelines do not carry the force of law.", "keyProvisions": [ "Rule 1.1 Comment [6] now requires lawyers to keep abreast of the benefits and risks associated with existing and developing technology", "Lawyers must review all GAI output for accuracy, verify citations and facts, and correct false or misleading arguments before filing", "A lawyer may charge only for actual time spent; everyday GAI tool costs are non-billable overhead, and GAI training costs generally may not be billed", "A lawyer must disclose GAI use when the client asks, or when entering client information into a GAI system", "Judicial officers may consult but should never rely on GAI-produced content when drafting orders and decisions, and GAI must never influence or appear to influence a conclusion", "Rule 2.9(C) risk: GAI output may import outside information amounting to prohibited independent factual investigation", "Even a single incorrect citation may trigger a Rule 8.3 reporting obligation, including as to opposing counsel's output" ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court order and both sets of interim guidelines (official PDF)", "url": "https://www.courts.ri.gov/Miscellaneous%20Orders/Supreme-2026-06_In_reAmendments_Article_V_Rule_1_1_Professional_Conduct.pdf" }, { "label": "Interim Guidelines … by Rhode Island Lawyers (official PDF)", "url": "https://www.courts.ri.gov/attorney-resources/Documents/Interim_Guidelines_Ethical_Use_of_Generative_AI_by_RI_Lawyers.pdf" } ], "country": "US", "categories": [ "bar-guidance", "judicial-use", "verification-duty", "confidentiality-client-data", "fees-billing", "competence-duty", "supervision", "evidence-admissibility", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ec-comite-etica-funcion-judicial-protocolo-ia-2026", "url": "https://safelegalai.com/regulation/other/ec-comite-etica-funcion-judicial-protocolo-ia-2026", "title": "Protocol for the use of artificial intelligence in the Judicial Function (Ethics Committee)", "body": "Ethics Committee of the Judicial Function (Comite de Etica de la Funcion Judicial)", "bodyShort": "Comite de Etica FJ", "jurisdiction": "other", "type": "policy", "status": "proposed", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "disclosure", "confidentiality" ], "versions": [ { "version": "Approved by the Ethics Committee; formal Judiciary Council resolution pending", "date": "2026-06-16T00:00:00.000Z", "url": "https://www.eluniverso.com/noticias/politica/funcion-judicial-de-ecuador-avanza-en-la-elaboracion-de-un-protocolo-para-el-uso-de-la-inteligencia-artificial-nota/", "changes": "Initial protocol; no official URL located, content from El Universo reporting." } ], "summary": "Twenty-three-principle protocol drafted by the Judiciary Council, National Court, Prosecutor General and Public Defender: no AI may issue, replace or determine judicial decisions; public tools are barred for case data and only institutional tools allowed; independent verification is mandatory; every procedural document must declare when and how AI was used. It awaits formal adoption.", "keyProvisions": [ "No AI may issue, replace or determine judicial decisions; judges may not motivate rulings with algorithms or assess credibility via AI.", "Public AI tools prohibited for case data; institutional tools only.", "Independent verification of AI information before use.", "Declaration of AI use in all procedural documents.", "No AI translation or analysis of Kichwa and Shuar testimony." ], "relatedIncidents": [], "sources": [ { "label": "El Universo report", "url": "https://www.eluniverso.com/noticias/politica/funcion-judicial-de-ecuador-avanza-en-la-elaboracion-de-un-protocolo-para-el-uso-de-la-inteligencia-artificial-nota/" } ], "country": "EC", "categories": [ "judicial-use", "ai-decision-prohibition", "disclosure-filings", "confidentiality-client-data", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "br-oab-plano-nacional-ia-advocacia-2026", "url": "https://safelegalai.com/regulation/brazil/br-oab-plano-nacional-ia-advocacia-2026", "title": "National Plan for Integrating Artificial Intelligence into Legal Practice", "body": "Federal Council of the Brazilian Bar (Conselho Federal da OAB)", "bodyShort": "OAB", "jurisdiction": "brazil", "type": "policy", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "permissive" ], "effectiveDate": "2026-06-15T00:00:00.000Z", "versions": [ { "version": "Launched at the Conselho Pleno, Joao Pessoa", "date": "2026-06-15T00:00:00.000Z", "url": "https://www.oab.org.br/noticia/64268/conselho-federal-da-oab-anuncia-plano-nacional-para-integrar-inteligencia-artificial-a-advocacia", "changes": "Elevates Recomendacao 001/2024 to institutional policy." } ], "summary": "Institutional plan launched on 15 June 2026 with five axes: AI governance and good practice, training of lawyers through the ESA schools, modernisation of OAB services, defence of professional prerogatives, and technological inclusion of small practices. It commits the Bar to a Code of Good Practice for AI in Legal Practice and to regional technical chambers.", "keyProvisions": [ "Code of Good Practice for AI in Legal Practice to be drafted and periodically updated.", "Expanded AI, data-protection and ethics coursework via the Superior Schools of Advocacy.", "Sectional technical chambers to adapt national guidance regionally.", "Not a binding rule; builds on Recomendacao 001/2024." ], "relatedIncidents": [], "sources": [ { "label": "OAB news: National AI Plan", "url": "https://www.oab.org.br/noticia/64268/conselho-federal-da-oab-anuncia-plano-nacional-para-integrar-inteligencia-artificial-a-advocacia" }, { "label": "Migalhas report", "url": "https://www.migalhas.com.br/quentes/458194/oab-lanca-plano-nacional-para-orientar-uso-de-ia-na-advocacia" } ], "country": "BR", "categories": [ "bar-guidance", "competence-duty", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "pl-nra-ethics-code-ai-amendments", "url": "https://safelegalai.com/regulation/other/pl-nra-ethics-code-ai-amendments", "title": "Amendments to the Collection of Rules of Advocates' Ethics and Dignity of the Profession introducing § 23e on technology and AI", "body": "Polish Bar Council [Naczelna Rada Adwokacka]", "bodyShort": "NRA", "jurisdiction": "other", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "disclosure" ], "versions": [ { "version": "Adopted at the NRA plenary session in Rzeszów", "date": "2026-06-13T00:00:00.000Z", "url": "https://www.adwokatura.pl/z-zycia-nra/zmiany-w-zbiorze-zasad-etyki-adwokackiej-naczelna-rada-adwokacka-przyjela-nowe-regulacje-dotyczace-korzystania-ze-sztucznej-inteligencji-w-wykonywaniu-zawodu-adwokata/", "changes": "New § 23e added and § 19(6) amended at the plenary session held 12–13 June 2026; resolution number not stated in the published announcement." } ], "summary": "A resolution of the Polish Bar Council adopted at its 12–13 June 2026 plenary in Rzeszów amending the advocates' ethics code. It adds § 23e regulating the use of technological tools, including AI, and amends § 19(6). The rules bind Polish advocates and are enforceable through the disciplinary system.", "keyProvisions": [ "An advocate may use technological tools, including AI, only in an auxiliary capacity.", "Technology may not process information covered by professional secrecy unless the tool's mechanisms guarantee proper protection of that data.", "The advocate must personally assess and verify the results produced by the tool so as not to harm the client.", "Use of technological tools does not release the advocate from personal responsibility for the form and content of professional activities.", "Clients need not generally be told which tools are used, but may request disclosure of the technologies used and their purpose.", "The advocate must have the competence to assess the tool's limitations and associated risks." ], "relatedIncidents": [], "sources": [ { "label": "Naczelna Rada Adwokacka — announcement of the ethics code amendments", "url": "https://www.adwokatura.pl/z-zycia-nra/zmiany-w-zbiorze-zasad-etyki-adwokackiej-naczelna-rada-adwokacka-przyjela-nowe-regulacje-dotyczace-korzystania-ze-sztucznej-inteligencji-w-wykonywaniu-zawodu-adwokata/" } ], "country": "PL", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "co-consejo-de-estado-ai-rules-2026", "url": "https://safelegalai.com/regulation/other/co-consejo-de-estado-ai-rules-2026", "title": "Council of State, Third Section: parameters for AI use in proceedings and disciplinary referral", "body": "Council of State (Consejo de Estado, Seccion Tercera)", "bodyShort": "Consejo de Estado", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "judges", "parties" ], "requirements": [ "verification" ], "effectiveDate": "2026-06-09T00:00:00.000Z", "versions": [ { "version": "Decision reported 9 June 2026", "date": "2026-06-09T00:00:00.000Z", "url": "https://www.infobae.com/colombia/2026/06/09/consejo-de-estado-fijo-reglas-para-usar-la-inteligencia-artificial-en-la-justicia-tras-caso-de-abogado-que-habria-inventado-sentencias-con-chatgpt/", "changes": "Initial decision; reported by Infobae and El Tiempo, official text not obtained." } ], "summary": "In an extraordinary unification appeal citing rulings the Court's reporting office could not locate, the Council of State held that AI may only assist with searching and organising information, that all AI output requires human verification, that adjudication is non-delegable, and that AI-detection tools are too unreliable to carry judicial weight. It referred the lawyer for discipline.", "keyProvisions": [ "AI is limited to search, organisation and processing of information.", "Human verification required before any AI output enters an official document.", "Judges may not delegate evidence evaluation, fact interpretation or decisions to AI.", "AI-detector output lacks the reliability needed for judicial findings.", "Referral to the Comision Seccional de Disciplina Judicial de Bogota." ], "relatedIncidents": [], "sources": [ { "label": "Infobae report", "url": "https://www.infobae.com/colombia/2026/06/09/consejo-de-estado-fijo-reglas-para-usar-la-inteligencia-artificial-en-la-justicia-tras-caso-de-abogado-que-habria-inventado-sentencias-con-chatgpt/" }, { "label": "El Tiempo report", "url": "https://www.eltiempo.com/justicia/cortes/el-duro-llamado-de-atencion-del-consejo-de-estado-sobre-el-uso-irresponsable-de-la-inteligencia-artificial-en-demandas-3561102" } ], "country": "CO", "categories": [ "sanctions-record", "verification-duty", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-fre-707-machine-generated-evidence", "url": "https://safelegalai.com/regulation/us-federal/us-fre-707-machine-generated-evidence", "title": "Proposed Federal Rule of Evidence 707: Machine-Generated Evidence", "body": "Judicial Conference of the United States, Committee on Rules of Practice and Procedure", "bodyShort": "US Judicial Conference", "jurisdiction": "us-federal", "type": "court-rule", "status": "proposed", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "verification" ], "versions": [ { "version": "Returned for further study (Standing Committee, June 2026)", "date": "2026-06-04T00:00:00.000Z", "url": "https://www.uscourts.gov/sites/default/files/document/2026-06_standing_committee_meeting_agenda_book_final.pdf" } ], "summary": "A proposed United States federal evidence rule that would subject machine-generated evidence, including AI output offered without an expert, to the reliability standards of Rule 702. Public comment closed 16 February 2026. In June 2026 the Standing Committee declined to recommend adoption and returned the rule for further study alongside deepfakes; the earliest effective date is after 1 December 2027.", "keyProvisions": [ "Would apply Rule 702 reliability requirements to evidence produced by a machine or process without a testifying expert.", "Public comment period ran to 16 February 2026.", "Standing Committee did not recommend action on 3 to 4 June 2026; rule revised and returned for study with deepfake issues.", "Reported in the September 2026 reports to the Judicial Conference." ], "relatedIncidents": [], "sources": [ { "label": "US Courts — Reports to the Judicial Conference, September 2026", "url": "https://www.uscourts.gov/forms-rules/records-rules-committees/committee-reports/reports-judicial-conference-september-2026" }, { "label": "US Courts — Standing Committee agenda book, June 2026 (PDF)", "url": "https://www.uscourts.gov/sites/default/files/document/2026-06_standing_committee_meeting_agenda_book_final.pdf" } ], "country": "US", "categories": [ "evidence-admissibility" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "india-supreme-court-draft-ai-regulations-2026", "url": "https://safelegalai.com/regulation/india/india-supreme-court-draft-ai-regulations-2026", "title": "Draft Regulations for Use of Artificial Intelligence in Courts, 2026", "body": "Supreme Court of India (AI Committee)", "bodyShort": "Supreme Court of India", "jurisdiction": "india", "type": "regulation", "status": "consultation", "appliesTo": [ "judges", "court-staff", "lawyers", "parties" ], "requirements": [ "disclosure", "prohibition", "verification", "record-keeping" ], "versions": [ { "version": "Draft for public comment", "date": "2026-06-03T00:00:00.000Z", "url": "https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/uploads/2026/06/2026060342.pdf" } ], "summary": "Draft regulations published by the Supreme Court of India's AI Committee on 3 June 2026 for the Supreme Court, all High Courts and all subordinate courts and tribunals. AI may assist with research, drafting, transcription and citation checks but may not decide cases or assess credibility; lawyers must disclose AI use in pleadings. Comments closed 15 July 2026.", "keyProvisions": [ "Applies to the Supreme Court, all High Courts, and all courts, tribunals and adjudicatory commissions.", "Prohibits AI from deciding cases, bail, witness credibility or influencing judicial deliberation.", "Lawyers must disclose the use of AI in pleadings.", "Permits AI for research, drafting assistance, transcription, translation, citation verification and case management.", "Creates institutional AI Committees, secretariats and a central CoRE-AI body." ], "relatedIncidents": [ "pooja-ramesh-singh-v-jk-bank" ], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "Supreme Court of India — draft AI regulations (PDF)", "url": "https://cdnbbsr.s3waas.gov.in/s3ec0490f1f4972d133619a60c30f3559e/uploads/2026/06/2026060342.pdf" } ], "country": "IN", "categories": [ "disclosure-filings", "judicial-use", "ai-decision-prohibition", "verification-duty", "court-own-use", "litigant-in-person", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-oh-ethics-guide-artificial-intelligence", "url": "https://safelegalai.com/regulation/us-state/us-oh-ethics-guide-artificial-intelligence", "title": "Ohio Ethics Guide: Artificial Intelligence for Lawyers and Judicial Officers", "body": "Ohio Board of Professional Conduct", "bodyShort": "Ohio Board of Professional Conduct", "jurisdiction": "us-state", "region": "Ohio", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "judges", "firms" ], "requirements": [ "competence", "confidentiality", "verification", "supervision" ], "versions": [ { "version": "Original", "date": "2026-06-02T00:00:00.000Z", "url": "https://kybar.org/Portals/0/Documents/Ohio%20AI%20Guide%20for%20Attorneys.pdf", "changes": "Issued and distributed to Ohio lawyers and judicial officers" } ], "summary": "Nonbinding Board staff guidance applying the Ohio Rules of Professional Conduct and the Code of Judicial Conduct to AI. It covers competence, confidentiality, client communication, independent professional judgment, fees, candor to the tribunal and hallucinated citations, supervision of staff and vendors, and a substantial judicial section on competence, independent decision-making, ex parte risk and confidentiality.", "keyProvisions": [ "The Ohio Rules of Professional Conduct do not prohibit lawyers from using AI, but AI is now a 'relevant technology' for the duty of competence", "Lawyers may bill only for time actually spent, not the time a task would have taken without AI; standard AI tool costs are usually non-billable overhead", "Filing documents with fictitious citations, quotes or propositions of law implicates Prof.Cond.R. 3.3(a), 8.4(c), 8.4(d) and 3.1", "A lawyer who learns a filing contains a hallucination must immediately notify the court and opposing counsel", "A lawyer must comply with local rules and standing orders on the use and disclosure of AI tools (Prof.Cond.R. 3.4)", "A judge should never use AI to decide the ultimate conclusion in a case and should avoid AI-produced first drafts of decisions or orders", "Judges and staff must not enter confidential case information, or anything revealing judicial mental processes, into general or free public AI tools" ], "relatedIncidents": [], "sources": [ { "label": "Ohio Ethics Guide (official PDF)", "url": "https://kybar.org/Portals/0/Documents/Ohio%20AI%20Guide%20for%20Attorneys.pdf" }, { "label": "Supreme Court of Ohio, AI Resource Library", "url": "https://www.supremecourt.ohio.gov/courts/services-to-courts/artificial-intelligence-resource-library/" }, { "label": "Court News Ohio (30 Jun 2026)", "url": "https://www.courtnewsohio.gov/happening/2026/AIEthics_063026.asp" } ], "country": "US", "categories": [ "bar-guidance", "judicial-use", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ee-aki-tehisaru-guidance", "url": "https://safelegalai.com/regulation/other/ee-aki-tehisaru-guidance", "title": "Artificial intelligence guidance hub [Tehisaru]", "body": "Estonian Data Protection Inspectorate [Andmekaitse Inspektsioon]", "bodyShort": "Andmekaitse Inspektsioon", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification", "competence", "record-keeping" ], "versions": [ { "version": "Current", "date": "2026-06-01T00:00:00.000Z", "url": "https://www.aki.ee/tehisaru/tehisaru-ja-andmekaitse/tehisintellekti-maarus", "changes": "Standing section covering AI and data protection, the AI Regulation, personal-data protection, awareness and human rights, and a page for AI adopters; supplemented in June 2026 by opinions on general-purpose AI and AI documentation software in healthcare, and by a Digital Omnibus / AI Omnibus timetable page. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "The Estonian DPA's standing guidance area on artificial intelligence, explaining the EU AI Act's risk-tiered structure, the interaction between AI and personal-data protection, individuals' rights, and practical duties for organisations adopting AI. It also tracks the Digital Omnibus and AI Omnibus legislative timetable.", "keyProvisions": [ "Explains that the AI Regulation does not treat AI as inherently dangerous but sets limits and duties on uses that materially affect people's employment, education, benefits, services, safety or privacy.", "Sets out the tiered classification from permitted low-risk uses through to uses prohibited in the EU, with obligations scaling to potential harm.", "A dedicated page for AI adopters [Tehisaru kasutusele võtjale] addresses deployer-side data-protection duties.", "A June 2026 response to a request for explanation addresses AI-based documentation software, and an accompanying notice addresses general-purpose AI use in healthcare." ], "relatedIncidents": [], "sources": [ { "label": "AKI — Tehisintellekti määrus", "url": "https://www.aki.ee/tehisaru/tehisaru-ja-andmekaitse/tehisintellekti-maarus" }, { "label": "AKI — Tehisaru kasutusele võtjale", "url": "https://www.aki.ee/tehisaru/tehisaru-ja-andmekaitse/tehisaru-kasutusele-votjale" } ], "country": "EE", "categories": [ "data-protection", "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "law-society-generative-ai-essentials", "url": "https://safelegalai.com/regulation/uk/law-society-generative-ai-essentials", "title": "Generative AI: the essentials", "body": "Law Society of England and Wales", "bodyShort": "Law Society", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence", "supervision" ], "effectiveDate": "2023-11-01T00:00:00.000Z", "versions": [ { "version": "June 2026 update", "date": "2026-06-01T00:00:00.000Z", "url": "https://www.lawsociety.org.uk/topics/ai-and-lawtech/generative-ai-the-essentials", "changes": "Refreshed guidance on choosing tools, human oversight and client confidentiality; remains the Law Society's current statement on generative AI for solicitors." }, { "version": "Original", "date": "2023-11-01T00:00:00.000Z", "url": "https://www.lawsociety.org.uk/topics/ai-and-lawtech/generative-ai-the-essentials" } ], "summary": "The Law Society of England and Wales' practical guide to generative AI for solicitors and firms, updated in June 2026. It is permissive with cautions: solicitors may use generative AI but must keep a human in the loop, verify output, protect client confidentiality, understand the tool's limits and ensure staff are competent and supervised.", "keyProvisions": [ "Generative AI output must be reviewed by a competent human before it is relied on.", "Client and confidential information should not be entered into tools without appropriate contractual and technical safeguards.", "Firms should set an AI use policy and train staff on approved tools.", "Solicitors should understand how the tool works, including whether prompts are retained or used for training." ], "relatedIncidents": [ "ayinde-v-haringey", "cork-v-smith" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "Law Society — Generative AI: the essentials", "url": "https://www.lawsociety.org.uk/topics/ai-and-lawtech/generative-ai-the-essentials" } ], "country": "GB", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "sra-effective-supervision-guidance", "url": "https://safelegalai.com/regulation/uk/sra-effective-supervision-guidance", "title": "Effective supervision guidance (AI sections)", "body": "Solicitors Regulation Authority", "bodyShort": "SRA", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "supervision", "competence", "verification" ], "effectiveDate": "2026-06-01T00:00:00.000Z", "versions": [ { "version": "June 2026 update (AI sections added)", "date": "2026-06-01T00:00:00.000Z", "url": "https://www.sra.org.uk/solicitors/guidance/effective-supervision-guidance/" } ], "summary": "The SRA's Effective Supervision guidance for solicitors and firms in England and Wales gained dedicated AI sections in June 2026. Supervisors are treated as accountable for AI-assisted work by staff and must have systems for supervising matters where AI is used. The 17 August 2026 warning notice on AI misuse cross-refers to it.", "keyProvisions": [ "Supervisors remain accountable for work product produced with AI assistance by the people they supervise.", "Firms need systems that let supervisors see how AI is being used on matters.", "Junior staff and support staff using AI tools must be supervised in the same way as any other delegated work.", "Referenced as a companion document in the SRA warning notice on misuse of AI." ], "relatedIncidents": [ "cork-v-smith" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "SRA — Effective supervision guidance", "url": "https://www.sra.org.uk/solicitors/guidance/effective-supervision-guidance/" }, { "label": "SRA press release referencing the June 2026 update", "url": "https://www.sra.org.uk/news/news/releases/responsible-use-ai/" } ], "country": "GB", "categories": [ "supervision", "competence-duty", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "us-ny-22-nycrr-part-161-artificial-intelligence", "url": "https://safelegalai.com/regulation/us-state/us-ny-22-nycrr-part-161-artificial-intelligence", "title": "22 NYCRR Part 161 — Use of Artificial Intelligence Technology", "body": "New York State Unified Court System (Rules of the Chief Administrator of the Courts)", "bodyShort": "NY Unified Court System", "jurisdiction": "us-state", "region": "New York", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "certification", "verification", "permissive" ], "effectiveDate": "2026-06-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-06-01T00:00:00.000Z", "url": "https://www.nycourts.gov/rules/part-161-use-artificial-intelligence-technology", "changes": "Added to the Rules of the Chief Administrator by Administrative Order AO/75/26 dated 25 March 2026; effective 1 June 2026" } ], "summary": "Part 161 applies to all New York Unified Court System courts in civil and criminal cases. Using an AI tool to prepare a paper is permitted and need not be disclosed, but the attorney or party must independently ensure it contains no fabricated cases, statutes or other material; signing certifies that review, and failure may draw sanctions.", "keyProvisions": [ "Applies to all UCS courts in both civil and criminal cases", "Defines AI as any machine-based system capable of making predictions, recommendations and/or decisions based on human-defined objectives and inputs", "Use of an AI tool to prepare a paper is permitted and need not be disclosed", "Any attorney or party who uses an AI tool must carefully review the paper and independently ensure it contains no fabricated or fictitious cases, statutes or other material", "Signing the paper certifies that such a review was conducted", "Failure to satisfy the requirement may result in sanction or other remedial action" ], "relatedIncidents": [], "sources": [ { "label": "22 NYCRR Part 161 (official rule page)", "url": "https://www.nycourts.gov/rules/part-161-use-artificial-intelligence-technology" }, { "label": "Administrative Order of the Chief Administrative Judge (official PDF)", "url": "https://www.nycourts.gov/LegacyPDFS/rules/comments/pdf/AdministrativeOrder-CAJ-75-2026-ArttificialIntelligence-032526r.pdf" }, { "label": "NYSBA notice on the new rule", "url": "https://nysba.org/effective-june-1-2026-the-new-york-state-unified-court-system-has-adopted-a-new-rule-regarding-the-use-of-artificial-intelligence/" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-fl-rule-2-515-generative-ai-amendment", "url": "https://safelegalai.com/regulation/us-state/us-fl-rule-2-515-generative-ai-amendment", "title": "In re Amendments to Florida Rule of General Practice and Judicial Administration 2.515 (generative AI), SC2026-0673", "body": "Supreme Court of Florida", "bodyShort": "Florida Supreme Court", "jurisdiction": "us-state", "region": "Florida", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "certification", "verification" ], "versions": [ { "version": "SC2026-0673", "date": "2026-05-28T00:00:00.000Z", "url": "https://flcourts-media.flcourts.gov/content/download/2489374/opinion/Opinion_SC2026-0673.pdf", "changes": "Amendment of rule 2.515(d)(2), rule retitled Signature and Representations to Court" } ], "summary": "Florida moved from bar guidance to a binding filing rule. The Supreme Court amended Rule of General Practice and Judicial Administration 2.515(d)(2) on 28 May 2026 to address generative AI tools in the representations a signer makes when filing a document, alongside conforming amendments to rules 2.505, 2.520 and 2.525.", "keyProvisions": [ "Rule 2.515 retitled Signature and Representations to Court", "Rule 2.515(d)(2) (Representation by Signer) amended to require the signer to account for generative AI-produced content", "Signer representations include that the signer has read the document", "Conforming amendments to rules 2.505, 2.520 and 2.525", "Operative wording not yet retrieved; must be confirmed from the official opinion before publication" ], "relatedIncidents": [], "sources": [ { "label": "Opinion SC2026-0673 (PDF)", "url": "https://flcourts-media.flcourts.gov/content/download/2489374/opinion/Opinion_SC2026-0673.pdf" }, { "label": "Opinion index entry", "url": "https://www.courtlistener.com/opinion/10865530/in-re-amendments-to-florida-rule-of-general-practice-and-judicial/" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "eu-commission-draft-guidelines-high-risk-classification", "url": "https://safelegalai.com/regulation/eu/eu-commission-draft-guidelines-high-risk-classification", "title": "Draft Commission Guidelines on the classification of high-risk AI systems (Article 6)", "body": "European Commission", "bodyShort": "European Commission", "jurisdiction": "eu", "type": "guidance", "status": "consultation", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Draft for targeted consultation", "date": "2026-05-19T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/library/draft-commission-guidelines-classification-high-risk-ai-systems" } ], "summary": "Draft guidelines published by the European Commission on 19 May 2026 on classifying AI systems as high-risk under Article 6 of the EU AI Act, including administration-of-justice uses listed in Annex III. A targeted consultation closed in summer 2026 and final guidelines are expected by the end of 2026, ahead of the deferred 2 December 2027 application date.", "keyProvisions": [ "Explains the Annex III use cases, including AI intended to assist judicial authorities in researching and interpreting law.", "Sets out the Article 6(3) derogation for systems that do not pose a significant risk.", "Draft only; final guidelines expected end of 2026." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — draft high-risk classification guidelines", "url": "https://digital-strategy.ec.europa.eu/en/library/draft-commission-guidelines-classification-high-risk-ai-systems" } ], "country": "EU", "categories": [ "ai-act-horizontal", "judicial-use" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "uk-bsb-ai-guidance", "url": "https://safelegalai.com/regulation/uk/uk-bsb-ai-guidance", "title": "Guidance on the use of Artificial Intelligence and Other Technologies", "body": "Bar Standards Board", "bodyShort": "BSB", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "barristers" ], "requirements": [ "verification", "disclosure", "confidentiality", "competence" ], "effectiveDate": "2026-05-18T00:00:00.000Z", "versions": [ { "version": "May 2026", "date": "2026-05-18T00:00:00.000Z", "url": "https://www.barstandardsboard.org.uk/static/5e1baa5c-614c-4105-afae9a9ccdf8d97b/c0d1ed62-fcad-4a68-8ab810de5fa9b116/Artificial-Intelligence-Guidance-May-2026.pdf" } ], "summary": "Regulatory guidance from the Bar Standards Board, valid from 18 May 2026, on how the Core Duties apply when barristers in England and Wales use AI. It requires verification of output before use, transparency with clients under rC19 where AI materially affects the service, and treats free public tools as generally unsuitable for legal work.", "keyProvisions": [ "Reads AI use into Core Duties CD1, CD2, CD3, CD6, CD7 and CD10.", "Barristers must verify AI output before relying on it or presenting it to a court.", "Where AI materially impacts the service provided, clients must be informed under rC19.", "Free consumer tools are generally unsuitable for confidential legal work." ], "relatedIncidents": [ "ayinde-v-haringey" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "Bar Standards Board — AI guidance (May 2026 PDF)", "url": "https://www.barstandardsboard.org.uk/static/5e1baa5c-614c-4105-afae9a9ccdf8d97b/c0d1ed62-fcad-4a68-8ab810de5fa9b116/Artificial-Intelligence-Guidance-May-2026.pdf" } ], "country": "GB", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "client-consent-transparency", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "int-coe-framework-convention-ai-cets-225", "url": "https://safelegalai.com/regulation/international/int-coe-framework-convention-ai-cets-225", "title": "Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (CETS No. 225)", "body": "Council of Europe", "bodyShort": "CoE", "jurisdiction": "international", "type": "statute", "status": "proposed", "appliesTo": [ "all" ], "requirements": [ "risk-classification", "disclosure" ], "versions": [ { "version": "EU ratification", "date": "2026-05-15T00:00:00.000Z", "url": "https://www.coe.int/en/web/artificial-intelligence/-/european-union-ratifies-the-council-of-europe-framework-convention-on-artificial-intelligence", "changes": "First ratification. Treaty Office chart of 3 September 2026 shows 20 signatures without ratification; not yet in force." }, { "version": "Opened for signature", "date": "2024-09-05T00:00:00.000Z", "url": "https://www.coe.int/en/web/conventions/full-list?module=treaty-detail&treatynum=225", "changes": "Signed in Vilnius by 10 states plus the EU; corrected certified copy per Committee of Ministers decision of 30 April 2025." }, { "version": "Adopted by the Committee of Ministers", "date": "2024-05-17T00:00:00.000Z", "url": "https://rm.coe.int/1680afae3c" } ], "summary": "The first international treaty on AI, adopted 17 May 2024 and opened for signature in Vilnius on 5 September 2024. Parties must ensure AI lifecycle activities respect human rights, democracy and the rule of law, with transparency, accountability, remedies and risk assessment. As of 3 September 2026 only the EU has ratified; five ratifications are needed.", "keyProvisions": [ "Applies to activities within the lifecycle of AI systems by public authorities and, at parties' option, private actors (art. 3).", "Transparency and oversight measures, including identification of AI-generated content (art. 8).", "Accessible and effective remedies for AI-related human-rights violations (art. 14).", "Procedural safeguards, including notice that a person is interacting with an AI system rather than a human (art. 15).", "Risk and impact assessment framework for AI systems (art. 16)." ], "relatedIncidents": [], "sources": [ { "label": "Treaty Office — treaty detail (CETS 225)", "url": "https://www.coe.int/en/web/conventions/full-list?module=treaty-detail&treatynum=225" }, { "label": "Treaty Office — chart of signatures and ratifications", "url": "https://www.coe.int/en/web/conventions/full-list?module=signatures-by-treaty&treatynum=225" }, { "label": "Treaty text (PDF)", "url": "https://rm.coe.int/1680afae3c" } ], "country": "INT", "categories": [ "ai-act-horizontal", "data-protection", "evidence-admissibility" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "vic-sc-gen-25-ai-court-users", "url": "https://safelegalai.com/regulation/australia/vic-sc-gen-25-ai-court-users", "title": "Practice Note SC GEN 25: The Use of Artificial Intelligence by Court Users", "body": "Supreme Court of Victoria", "bodyShort": "Supreme Court of Victoria", "jurisdiction": "australia", "region": "Victoria", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure" ], "effectiveDate": "2026-05-14T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-05-14T00:00:00.000Z", "url": "https://www.supremecourt.vic.gov.au/areas/legal-resources/practice-notes/sc-gen-25-the-use-of-artificial-intelligence-by-court-users" } ], "summary": "Practice Note SC GEN 25 of the Supreme Court of Victoria, issued 14 May 2026, replaces the court's May 2024 guidelines for litigants. Court users must keep meaningful human control over AI-assisted work, may not verify one AI tool's output with another, must ensure sworn documents are in the deponent's own words, and may be directed to identify AI-produced portions.", "keyProvisions": [ "Meaningful human control is required over any AI-assisted material filed with the court.", "Checking one AI tool's output with another AI tool does not satisfy the verification duty.", "Affidavits and witness statements must be in the deponent's own words.", "The court may direct a party to identify which parts of a document were AI-produced." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Victoria — Practice Note SC GEN 25", "url": "https://www.supremecourt.vic.gov.au/areas/legal-resources/practice-notes/sc-gen-25-the-use-of-artificial-intelligence-by-court-users" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person", "evidence-admissibility" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "vic-supreme-court-judicial-ai-guidelines", "url": "https://safelegalai.com/regulation/australia/vic-supreme-court-judicial-ai-guidelines", "title": "Judicial Guidelines on the Use of Artificial Intelligence", "body": "Supreme Court of Victoria", "bodyShort": "Supreme Court of Victoria", "jurisdiction": "australia", "region": "Victoria", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality", "verification" ], "effectiveDate": "2026-05-14T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-05-14T00:00:00.000Z", "url": "https://www.supremecourt.vic.gov.au/news/new-practice-note-and-judicial-guidelines-on-the-use-of-artificial-intelligence" } ], "summary": "Guidelines issued by the Supreme Court of Victoria on 14 May 2026, alongside Practice Note SC GEN 25, governing judicial officers' and chambers staff use of AI. Judges may not use generative AI for decision-making or the substantive content of judgments, must protect confidential case material, and must verify any AI-assisted administrative output.", "keyProvisions": [ "Judicial officers may not use generative AI for decision-making or drafting the reasoning of judgments.", "Confidential and suppressed case material must not be entered into external AI tools.", "Any administrative use of AI by chambers must be verified by the judicial officer." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Victoria — news release on practice note and judicial guidelines", "url": "https://www.supremecourt.vic.gov.au/news/new-practice-note-and-judicial-guidelines-on-the-use-of-artificial-intelligence" } ], "country": "AU", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "au-fcfcoa-pd-ai", "url": "https://safelegalai.com/regulation/australia/au-fcfcoa-pd-ai", "title": "Practice Direction: Use of Artificial Intelligence (PD-AI)", "body": "Federal Circuit and Family Court of Australia", "bodyShort": "FCFCOA", "jurisdiction": "australia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure" ], "effectiveDate": "2026-05-01T00:00:00.000Z", "versions": [ { "version": "v1 (May 2026)", "date": "2026-05-01T00:00:00.000Z", "url": "https://www.fcfcoa.gov.au/pd/pd-ai" } ], "summary": "A practice direction issued by the Federal Circuit and Family Court of Australia in May 2026 as a companion to the Federal Court's GPN-AI. It sets expectations for lawyers and self-represented litigants using generative AI in family law and general federal law matters, including verification of AI-assisted content and readiness to explain AI use to the court on request.", "keyProvisions": [ "Practitioners and parties are responsible for verifying AI-assisted documents before filing.", "The court may ask how generative AI was used in preparing a document.", "Companion to Federal Court of Australia practice note GPN-AI." ], "relatedIncidents": [ "valu-v-minister-immigration", "dayal-fedcfamc2f-1166", "jne24-v-minister-immigration" ], "sources": [ { "label": "FCFCOA — Practice Direction PD-AI", "url": "https://www.fcfcoa.gov.au/pd/pd-ai" }, { "label": "FCFCOA — PD-AI v1 PDF (May 2026)", "url": "https://www.fcfcoa.gov.au/sites/default/files/2026-05/fcfcoa_pd_-_use_of_artificial_intelligence_-_pd-ai_-_0526v1.pdf" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "us-al-state-bar-formal-opinion-2026-01", "url": "https://safelegalai.com/regulation/us-state/us-al-state-bar-formal-opinion-2026-01", "title": "Alabama State Bar Formal Opinion 2026-01: Artificial Intelligence Use — Best Practices Under Existing Professional Conduct Rules", "body": "Alabama State Bar, Office of General Counsel", "bodyShort": "Alabama State Bar", "jurisdiction": "us-state", "region": "Alabama", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "consent" ], "versions": [ { "version": "FO 2026-01", "date": "2026-05-01T00:00:00.000Z", "url": "https://www.alabar.org/assets/2026/05/FO-2026-01.pdf", "changes": "Initial publication; the source gives a month only (published May 2026), so the day is recorded as the 1st" } ], "summary": "The Alabama State Bar's first formal opinion on AI. It applies ARPC 1.1, 1.4, 1.5, 1.6, 3.1, 3.3, 5.1 and 5.3 to generative and agentic AI, requires independent verification of AI citations before filing, prefers enterprise or closed AI systems, bars billing for AI time savings, and supplies a compliance checklist.", "keyProvisions": [ "Independently verify every legal citation generated by AI against an authoritative database before filing (ARPC 1.1, 3.1, 3.3)", "Input into most generative AI platforms is disclosure; prefer enterprise or closed systems, anonymise prompts, and obtain informed consent where risk cannot be eliminated (ARPC 1.6)", "No Alabama rule mandates disclosure of AI use to clients in all circumstances, but proactive communication is prudent (ARPC 1.4)", "A lawyer may bill only for time actually spent; review and judgment time is billable, time saved by AI is not (ARPC 1.5)", "AI is to be supervised like a nonlawyer assistant; the supervising lawyer remains fully responsible for the work product (ARPC 5.1, 5.3)", "Check each court's local rules and standing orders for AI disclosure or certification requirements before filing" ], "relatedIncidents": [], "sources": [ { "label": "Formal Opinion 2026-01 (PDF)", "url": "https://www.alabar.org/assets/2026/05/FO-2026-01.pdf" }, { "label": "Landing page", "url": "https://www.alabar.org/office-of-general-counsel/formal-opinions/artificial-intelligence-use-best-practices-under-existing-professional-conduct-rules/" } ], "country": "US", "categories": [ "bar-guidance", "disclosure-filings", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pk-njpmc-national-ai-guidelines-judicial-institutions-2026", "url": "https://safelegalai.com/regulation/other/pk-njpmc-national-ai-guidelines-judicial-institutions-2026", "title": "National Guidelines for the Use of Artificial Intelligence in Judicial Institutions", "body": "National Judicial (Policy Making) Committee, drafted by the National Judicial Automation Committee", "bodyShort": "NJPMC", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality", "competence" ], "effectiveDate": "2026-04-29T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-29T00:00:00.000Z", "url": "https://www.dawn.com/news/1996074/njpmc-issues-guidelines-for-ai-use-in-judicial-institutions", "changes": "Approved at the 57th NJPMC meeting; official text not located online." } ], "summary": "National guidelines approved by the NJPMC at its 57th meeting on 29 April 2026, developed by the National Judicial Automation Committee after consultation with all High Courts. AI is to assist, not replace, judicial decision-making; permitted uses are case management, legal research, predictive analytics and document processing, with bias, privacy and training safeguards.", "keyProvisions": [ "Human-centric: judges remain the ultimate arbiters and AI assists only.", "Permitted: case management, legal research, predictive analytics and document processing.", "Ethical and transparent use with safeguards against bias, plus explainability and accountability.", "Strict privacy and data-security standards.", "Structured training programmes for judges and court staff; High Courts retain implementation autonomy." ], "relatedIncidents": [], "sources": [ { "label": "Dawn report", "url": "https://www.dawn.com/news/1996074/njpmc-issues-guidelines-for-ai-use-in-judicial-institutions" }, { "label": "Business Recorder report", "url": "https://www.brecorder.com/news/amp/40418870" }, { "label": "Geo News report", "url": "https://www.geo.tv/latest/662137-njpmc-unveils-national-guidelines-on-use-of-ai-in-courts" } ], "country": "PK", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "training-cle", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "gh-national-ai-strategy-2025-2035", "url": "https://safelegalai.com/regulation/other/gh-national-ai-strategy-2025-2035", "title": "Ghana National Artificial Intelligence Strategy 2025-2035", "body": "Ministry of Communication, Digital Technology and Innovations", "bodyShort": "MoCDTI", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "permissive" ], "versions": [ { "version": "Launched by President John Dramani Mahama", "date": "2026-04-24T00:00:00.000Z", "url": "https://moc.gov.gh/2026/04/24/ghana-launches-national-ai-strategy-to-drive-digital-transformation-and-economic-growth/", "changes": "First national AI strategy for Ghana; announced alongside a proposed USD 250m AI computing centre. The strategy document itself was not obtained; the record rests on ministry launch announcements, so it is unverified." } ], "summary": "Ghana's first national AI strategy, launched 24 April 2026 by the Ministry of Communication, Digital Technology and Innovations. It targets education, healthcare, agriculture, finance and public service delivery, and proposes an independent Responsible AI Authority within the first year. Justice, courts and legal services are not named as target sectors.", "keyProvisions": [ "Framework built on multiple pillars with policy recommendations across eight target sectors; the administration of justice is not among them.", "Proposes an independent Responsible AI Authority modelled on comparable national AI offices.", "Implementation coordinated with the National Information Technology Agency, the Data Protection Commission and the Cyber Security Authority.", "Non-binding policy; Ghana has no AI statute and no AI risk-classification law." ], "relatedIncidents": [], "sources": [ { "label": "Ghana launches National AI Strategy (Ministry of Communication, Digital Technology and Innovations, 24 Apr 2026)", "url": "https://moc.gov.gh/2026/04/24/ghana-launches-national-ai-strategy-to-drive-digital-transformation-and-economic-growth/" }, { "label": "President Mahama launches Ghana's National AI Strategy (Ministry of the Interior, official)", "url": "https://www.mint.gov.gh/president-mahama-launches-ghanas-national-ai-strategy-calls-for-ai-that-reflects-ghanaian-values/" } ], "country": "GH", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "cl-corte-suprema-rol-23322-2025-ai-citations", "url": "https://safelegalai.com/regulation/other/cl-corte-suprema-rol-23322-2025-ai-citations", "title": "Supreme Court Rol 23.322-2025: suspension and fine for AI-fabricated doctrinal citations", "body": "Supreme Court of Chile, Third Chamber (Corte Suprema, Tercera Sala)", "bodyShort": "Corte Suprema", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "verification" ], "effectiveDate": "2026-04-22T00:00:00.000Z", "versions": [ { "version": "Decision of 22 April 2026", "date": "2026-04-22T00:00:00.000Z", "url": "https://www.pjud.cl/prensa-y-comunicaciones/noticias-del-poder-judicial/145652", "changes": "Initial decision; reported on the judiciary's own news service." } ], "summary": "The Supreme Court suspended a lawyer for one month and fined her 5 UTM under articles 531 and 542 of the Organic Code of Courts for a cassation brief citing a non-existent 2023 treatise and a non-existent 2021 work, produced with AI; the unintentional-error defence was rejected as a breach of procedural good faith.", "keyProvisions": [ "One-month suspension from practice plus a 5 UTM fine.", "Disciplinary basis: Codigo Organico de Tribunales arts 531 and 542.", "Lawyers must verify the accuracy of AI-generated sources; the good-faith excuse was rejected.", "Lower courts apply art. 531 similarly (2nd Civil Court of Concepcion, 1 UTM, February 2026)." ], "relatedIncidents": [], "sources": [ { "label": "Poder Judicial news", "url": "https://www.pjud.cl/prensa-y-comunicaciones/noticias-del-poder-judicial/145652" }, { "label": "24horas report", "url": "https://www.24horas.cl/actualidad/nacional/corte-suprema-suspende-abogada-citas-falsas-inteligencia-artificial" } ], "country": "CL", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "tw-judicial-yuan-ai-system-reference-guidelines-2026", "url": "https://safelegalai.com/regulation/other/tw-judicial-yuan-ai-system-reference-guidelines-2026", "title": "Reference Guidelines for the Use of Artificial Intelligence Systems by the Judicial Yuan and Subordinate Agencies", "body": "Judicial Yuan", "bodyShort": "Judicial Yuan", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "prohibition" ], "effectiveDate": "2026-04-22T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-22T00:00:00.000Z", "url": "https://lawplayer.com/article/6a08db0a3521bb47cdb12c8f", "changes": "Issued by order of 22 April 2026; the official page on judicial.gov.tw was not located, so a mirror of the national law database is cited." } ], "summary": "Judicial Yuan guidelines issued by order of 22 April 2026 for courts and judicial agencies using AI systems. Chapter 2 governs user conduct: users must understand that generative AI may hallucinate and must cross-check outputs against official documents, case files and legal databases before use; AI may only assist adjudication and deployments must incorporate risk management.", "keyProvisions": [ "Art. 8: verify AI-generated data, factual summaries and legal citations against reliable sources before use.", "AI assists adjudication only; users retain decision autonomy.", "Risk management required for system deployment." ], "relatedIncidents": [], "sources": [ { "label": "Guideline text (LawPlayer mirror of the national law database)", "url": "https://lawplayer.com/article/6a08db0a3521bb47cdb12c8f" }, { "label": "Judicial Weekly No. 2310 referencing the guideline (official PDF)", "url": "https://www.judicial.gov.tw/tw/dl-265614-a4837ab9748f48c8b534c45251be1e5b.html" } ], "country": "TW", "categories": [ "judicial-use", "verification-duty", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "au-federal-court-gpn-ai", "url": "https://safelegalai.com/regulation/australia/au-federal-court-gpn-ai", "title": "Generative Artificial Intelligence Practice Note (GPN-AI)", "body": "Federal Court of Australia", "bodyShort": "Federal Court of Australia", "jurisdiction": "australia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure" ], "effectiveDate": "2026-04-16T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-16T00:00:00.000Z", "url": "https://www.fedcourt.gov.au/law-and-practice/practice-documents/practice-notes/gpn-ai" } ], "summary": "Practice note GPN-AI, issued by the Federal Court of Australia on 16 April 2026, sets the court's expectations of all participants who use generative AI in proceedings. Practitioners and parties remain responsible for verifying AI-assisted content, must be able to explain how AI was used if asked, and the court may make orders regulating AI use in a particular proceeding.", "keyProvisions": [ "Parties and practitioners are responsible for the accuracy of any AI-assisted material filed.", "The court may direct a party to disclose whether and how generative AI was used.", "The court may make case-specific orders limiting or conditioning AI use.", "Affidavit and expert evidence must reflect the deponent's own knowledge and opinion." ], "relatedIncidents": [ "valu-v-minister-immigration", "dayal-fedcfamc2f-1166", "jne24-v-minister-immigration" ], "sources": [ { "label": "Federal Court of Australia — GPN-AI", "url": "https://www.fedcourt.gov.au/law-and-practice/practice-documents/practice-notes/gpn-ai" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "br-oab-sp-ted-supervisao-ia-2026", "url": "https://safelegalai.com/regulation/brazil/br-oab-sp-ted-supervisao-ia-2026", "title": "OAB/SP Ethics and Discipline Tribunal ruling on partners' duty to supervise AI use (proc. 25.0886.2025.014989-0)", "body": "Sao Paulo Bar Ethics and Discipline Tribunal (Tribunal de Etica e Disciplina da OAB/SP, 1a Turma)", "bodyShort": "OAB/SP TED", "jurisdiction": "brazil", "region": "Sao Paulo", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "supervision", "verification" ], "effectiveDate": "2026-04-16T00:00:00.000Z", "versions": [ { "version": "Ruling of 16 April 2026 (published May 2026)", "date": "2026-04-16T00:00:00.000Z", "url": "https://www.migalhas.com.br/quentes/455537/advogados-socios-devem-garantir-supervisao-em-uso-de-ia-decide-oab-sp", "changes": "Initial ruling; reported by Migalhas, official text not obtained." } ], "summary": "The Sao Paulo Bar's ethics tribunal held that partners and managers must ensure AI use by associates, contractors, trainees and non-lawyer staff is supervised, and must fully review all AI-generated output before it is filed, with special care for case-law research, citing article 77 of the Code of Civil Procedure and OAB Recommendation 001/2024.", "keyProvisions": [ "Partners are responsible for supervising AI use across the firm, including by non-lawyers.", "Full review of every AI output before filing.", "Heightened care where AI is used for legal research and case-law.", "Applies CPC art. 77 duty of truthfulness and Recomendacao 001/2024." ], "relatedIncidents": [], "sources": [ { "label": "Migalhas report", "url": "https://www.migalhas.com.br/quentes/455537/advogados-socios-devem-garantir-supervisao-em-uso-de-ia-decide-oab-sp" } ], "country": "BR", "categories": [ "supervision", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "es-cgae-circular-3-2026-ia-generativa-verificacion", "url": "https://safelegalai.com/regulation/other/es-cgae-circular-3-2026-ia-generativa-verificacion", "title": "CGAE Interpretive Circular 3/2026: use of generative AI systems and the duty of verification", "body": "General Council of Spanish Lawyers (Consejo General de la Abogacia Espanola)", "bodyShort": "CGAE", "jurisdiction": "other", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "supervision", "competence", "record-keeping" ], "effectiveDate": "2026-04-13T00:00:00.000Z", "versions": [ { "version": "Dated Madrid 20 March 2026; approved by the Pleno 10 April 2026, amended 13 April 2026", "date": "2026-04-13T00:00:00.000Z", "url": "https://www.icaib.org/wp-content/uploads/2026/04/Circular-interpretativa-3-2026_Deontologia-aprobada-Pleno-modificaciones-13-4-2026.pdf", "changes": "Original text, from the ICAIB mirror of the CGAE circular." } ], "summary": "Interpretive circular under Article 23 of the Right of Defence Act construing the Spanish code of conduct. Generative-AI drafting is lawful, but filing AI-assisted documents with unverified errors breaches diligence and responsible technology use and is a serious infraction under Article 125.u EGAE; the signer is solely responsible and tool failure proves a breach of supervision.", "keyProvisions": [ "AI is an auxiliary function subject to human supervision, never a substitute for the lawyer.", "Uncritical or substitutive delegation of essential professional tasks breaches diligence and independence, with or without errors.", "Serious infraction under Art. 125.u EGAE; very serious under Art. 124.g in extreme or repeated cases causing defencelessness.", "Recommendations: know the tools, read every AI output critically, cross-check against reliable legal sources, use AI only in fields mastered, keep internal traceability of AI use.", "Confidentiality and professional secrecy are reserved to a separate circular." ], "relatedIncidents": [], "sources": [ { "label": "Circular 3/2026 text (ICAIB mirror of the CGAE circular)", "url": "https://www.icaib.org/wp-content/uploads/2026/04/Circular-interpretativa-3-2026_Deontologia-aprobada-Pleno-modificaciones-13-4-2026.pdf" }, { "label": "CGAE news release", "url": "https://www.abogacia.es/actualidad/noticias/la-abogacia-espanola-advierte-de-los-riesgos-de-usar-la-ia-generativa-en-textos-profesionales/" } ], "country": "ES", "categories": [ "verification-duty", "supervision", "competence-duty", "bar-guidance", "sanctions-record", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "za-draft-national-ai-policy-withdrawn", "url": "https://safelegalai.com/regulation/south-africa/za-draft-national-ai-policy-withdrawn", "title": "Draft South Africa National Artificial Intelligence (AI) Policy (published GN 3880 of 2026; withdrawn by Notice 3978 of 2026)", "body": "Department of Communications and Digital Technologies", "bodyShort": "DCDT", "jurisdiction": "south-africa", "type": "policy", "status": "withdrawn", "appliesTo": [ "all", "providers" ], "requirements": [ "risk-classification", "disclosure" ], "versions": [ { "version": "Draft for public comment", "date": "2026-04-10T00:00:00.000Z", "url": "https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf", "changes": "Approved by Cabinet 25 March 2026 (and a special sitting on 1 April 2026); published as General Notice 3880 in Government Gazette 54477 for comment to 10 June 2026." }, { "version": "Withdrawal", "date": "2026-06-12T00:00:00.000Z", "url": "https://www.gov.za/sites/default/files/gcis_document/202606/54840gen3978.pdf", "changes": "Withdrawn in its entirety by Notice 3978 of 2026, Government Gazette 54840 (signed by Minister Solly Malatsi on 4 June 2026), after the draft was found to contain fabricated academic citations generated by AI." } ], "summary": "South Africa's draft national AI policy, gazetted for comment on 10 April 2026, proposed new oversight bodies including a National AI Commission. It was withdrawn in its entirety on 12 June 2026 after reporting established that several of its 67 academic citations — journals and authors alike — did not exist, apparently AI hallucinations in the policy's own drafting.", "keyProvisions": [ "Published as General Notice 3880 in Government Gazette 54477 on 10 April 2026 for public comment until 10 June 2026.", "Withdrawn in its entirety by Notice 3978 of 2026 in Government Gazette 54840, dated 12 June 2026 and signed 4 June 2026 by the Minister of Communications and Digital Technologies.", "Cabinet approved the withdrawal to allow a rework establishing national standards on the ethical use of AI.", "The withdrawal is the clearest South African precedent that AI-fabricated citations invalidate an official document." ], "relatedIncidents": [], "sources": [ { "label": "Withdrawal notice — Notice 3978 of 2026, GG 54840 (official gazette PDF)", "url": "https://www.gov.za/sites/default/files/gcis_document/202606/54840gen3978.pdf" }, { "label": "Draft policy — GN 3880, GG 54477 (official gazette PDF)", "url": "https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf" }, { "label": "SAnews — Minister announces withdrawal of draft AI Policy (official)", "url": "https://www.sanews.gov.za/south-africa/minister-announces-withdrawal-draft-ai-policy" } ], "country": "ZA", "categories": [ "ai-act-horizontal", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pt-csm-recomendacoes-uso-ia-atividade-jurisdicional", "url": "https://safelegalai.com/regulation/other/pt-csm-recomendacoes-uso-ia-atividade-jurisdicional", "title": "CSM Recommendations on the use of artificial intelligence in judicial activity (Recomendacoes no uso de IA)", "body": "High Council of the Judiciary (Conselho Superior da Magistratura)", "bodyShort": "CSM", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "disclosure", "verification", "prohibition", "confidentiality", "supervision" ], "effectiveDate": "2026-04-08T00:00:00.000Z", "versions": [ { "version": "Approved by the Plenario 8 April 2026; PDF published May 2026", "date": "2026-04-08T00:00:00.000Z", "url": "https://csm.org.pt/wp-content/uploads/2026/05/Recomendacoes_Uso_IA.pdf" } ], "summary": "Recommendations proposed by the CSM's AI Monitoring Group for judges' use of AI, citing AI Act Annex III point 8(a). Ten principles include effective human control, non-substitution, exclusive judicial responsibility and transparency; only systems supplied by the case-management platform or the CSM may be used; every AI-assisted decision must carry an express reference to AI use.", "keyProvisions": [ "Principle 2 and Rec. 7: AI may never replace judges in deciding, assessing facts or evidence, or interpreting the law.", "Rec. 6: complete, critical personal review and validation of any AI draft before it becomes a decision.", "Rec. 8: no judicial data in non-validated systems; open tools only for study on open sources.", "Rec. 10: express reference in the decision that it was produced with the aid of an AI system and fully reviewed by the signatory.", "Rec. 9: GDPR principles, security, traceability and no reuse; CSM-validated systems presumed compliant." ], "relatedIncidents": [], "sources": [ { "label": "CSM — Recomendacoes no uso de IA (PDF)", "url": "https://csm.org.pt/wp-content/uploads/2026/05/Recomendacoes_Uso_IA.pdf" }, { "label": "CSM announcement", "url": "https://csm.org.pt/csm-aprova-recomendacoes-para-o-uso-de-inteligencia-artificial-na-atividade-judicial/" } ], "country": "PT", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "data-protection", "training-cle", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "in-punjab-haryana-hc-ai-ban-judicial-officers", "url": "https://safelegalai.com/regulation/india/in-punjab-haryana-hc-ai-ban-judicial-officers", "title": "Directive prohibiting judicial officers from using AI tools for judgments and legal research", "body": "High Court of Punjab and Haryana (Registrar General, on direction of the Chief Justice)", "bodyShort": "Punjab and Haryana High Court", "jurisdiction": "india", "region": "Punjab, Haryana and Chandigarh", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition" ], "effectiveDate": "2026-04-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-06T00:00:00.000Z", "url": "https://lawtrend.in/punjab-and-haryana-high-court-bans-judicial-officers-from-using-ai-for-judgments-and-legal-research/", "changes": "Administrative communication reported 7 April 2026; official text not located." } ], "summary": "Administrative communication from the Punjab and Haryana High Court Registrar General, reported 7 April 2026, prohibiting all judicial officers in Punjab, Haryana and Chandigarh from using AI tools such as ChatGPT, Gemini, Copilot and Meta AI for drafting orders or judgments or for legal research, in order to preserve human judicial reasoning.", "keyProvisions": [ "No use of AI platforms for writing judgments or orders.", "No use of AI platforms for legal research.", "District and sessions judges responsible for compliance; breaches viewed seriously." ], "relatedIncidents": [], "sources": [ { "label": "Law Trend report", "url": "https://lawtrend.in/punjab-and-haryana-high-court-bans-judicial-officers-from-using-ai-for-judgments-and-legal-research/" } ], "country": "IN", "categories": [ "judicial-use", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "es-tsj-navarra-2026-citas-inventadas-sin-sancion", "url": "https://safelegalai.com/regulation/other/es-tsj-navarra-2026-citas-inventadas-sin-sancion", "title": "TSJ Navarra, Social Chamber (Apr 2026): Art. 247 LEC proceedings archived after apology for AI-invented citations", "body": "High Court of Justice of Navarre (Tribunal Superior de Justicia de Navarra)", "bodyShort": "TSJ Navarra", "jurisdiction": "other", "region": "Navarre", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "April 2026 (rapporteur Maria Jose Ramo)", "date": "2026-04-04T00:00:00.000Z", "url": "https://www.elespanol.com/espana/tribunales/20260404/abogada-cito-recurso-sentencias-inventadas-ia-libra-sancion-pedir-perdon-tribunal/1003744194050_0.html", "changes": "Reported by El Espanol on 4 April 2026; official text not obtained." } ], "summary": "Counsel's October 2025 appeal quoted passages attributed to the Constitutional Court, Supreme Court and two regional high courts that did not match any real ruling. The chamber opened Article 247 LEC bad-faith proceedings in February 2026 but archived them without fine after the lawyer apologised and withdrew the citations, while warning about AI hallucination and the non-delegable verification duty.", "keyProvisions": [ "Careless use of AI in judicial filings has deontological implications and can amount to procedural bad faith.", "Prompt acknowledgement and withdrawal weighed against sanction.", "Follows TSJ Navarra Auto 2/2024, the first Spanish decision on ChatGPT-generated filings." ], "relatedIncidents": [], "sources": [ { "label": "El Espanol (secondary)", "url": "https://www.elespanol.com/espana/tribunales/20260404/abogada-cito-recurso-sentencias-inventadas-ia-libra-sancion-pedir-perdon-tribunal/1003744194050_0.html" }, { "label": "Confilegal on the 2024 Navarra decision (secondary)", "url": "https://confilegal.com/20241115-el-tsj-de-navarra-decide-no-sancionar-a-un-abogado-que-presento-una-querella-con-la-ley-colombiana-al-usar-chatgpt/" } ], "country": "ES", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "sg-law-society-advisory-public-ai-tools", "url": "https://safelegalai.com/regulation/singapore/sg-law-society-advisory-public-ai-tools", "title": "Advisory on the Use of Publicly Available AI Tools", "body": "Law Society of Singapore", "bodyShort": "Law Society of Singapore", "jurisdiction": "singapore", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "verification", "competence" ], "effectiveDate": "2026-04-02T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-02T00:00:00.000Z", "url": "https://www.lawsociety.org.sg/wp-content/uploads/2026/04/Law-Societys-Advisory-on-the-Use-of-Publicly-Available-AI-Tools-2-April-2026.pdf", "archiveUrl": "https://web.archive.org/web/20260716080308/https://www.lawsociety.org.sg/wp-content/uploads/2026/04/Law-Societys-Advisory-on-the-Use-of-Publicly-Available-AI-Tools-2-April-2026.pdf" } ], "summary": "An advisory issued by the Law Society of Singapore on 2 April 2026 on lawyers' use of publicly available AI tools such as consumer chatbots. Client information must not be entered into tools that may retain or train on it, output must be verified against primary sources, and lawyers remain responsible under the Legal Profession rules.", "keyProvisions": [ "Client confidential information must not be entered into public AI tools without safeguards.", "Output from public tools must be verified against primary legal sources before use.", "Lawyers remain professionally responsible for AI-assisted work." ], "relatedIncidents": [ "tan-hai-peng-v-tan-cheong-joo" ], "article": "singapore-ai-guidance-history", "sources": [ { "label": "Law Society of Singapore — advisory on publicly available AI tools (PDF)", "url": "https://www.lawsociety.org.sg/wp-content/uploads/2026/04/Law-Societys-Advisory-on-the-Use-of-Publicly-Available-AI-Tools-2-April-2026.pdf" } ], "country": "SG", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "competence-duty", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-md-fla-berger-ai-standing-order", "url": "https://safelegalai.com/regulation/us-federal/us-md-fla-berger-ai-standing-order", "title": "Standing Order on Artificial Intelligence (Judge Wendy W. Berger)", "body": "United States District Court for the Middle District of Florida, Judge Wendy W. Berger", "bodyShort": "M.D. Fla. (Berger)", "jurisdiction": "us-federal", "region": "Florida", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "verification", "certification" ], "effectiveDate": "2026-04-02T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-02T00:00:00.000Z", "url": "https://www.flmd.uscourts.gov/sites/flmd/files/documents/flmd-berger-standing-order-on-ai.pdf" } ], "summary": "A standing order issued 2 April 2026 by Judge Wendy Berger of the Middle District of Florida governing AI-assisted filings. Counsel and self-represented parties are responsible for the accuracy of any AI-generated content; misrepresenting its accuracy can lead to sanctions up to dismissal or default judgment and referral to the state bar or the court's grievance committee.", "keyProvisions": [ "Filers remain fully responsible for verifying AI-generated citations, quotations and facts.", "Misrepresentation of the accuracy of AI content may result in dismissal, default or monetary sanctions.", "Attorneys may be referred to the Florida Bar or the court's grievance committee." ], "relatedIncidents": [], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "M.D. Fla. — Judge Berger standing order on AI (PDF)", "url": "https://www.flmd.uscourts.gov/sites/flmd/files/documents/flmd-berger-standing-order-on-ai.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "in-gujarat-hc-ai-policy-judicial-court-administration", "url": "https://safelegalai.com/regulation/india/in-gujarat-hc-ai-policy-judicial-court-administration", "title": "Policy on the Use of Artificial Intelligence in Judicial and Court Administration", "body": "High Court of Gujarat", "bodyShort": "Gujarat High Court", "jurisdiction": "india", "region": "Gujarat", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "confidentiality" ], "effectiveDate": "2026-04-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-04-01T00:00:00.000Z", "url": "https://gujarathighcourt.nic.in/hccms/sites/default/files/miscnotifications/Policy%20on%20the%20use%20of%20Artificial%20Intelligence%20in%20the%20Judicial%20and%20Court%20Administration.pdf", "changes": "Original policy; PDF metadata date, published April 2026." } ], "summary": "Gujarat High Court policy of April 2026 for the district judiciary, issued under Articles 227 and 235. AI is permitted solely as a neutral administrative and research instrument through High Court-defined tools; it may not be used for decision-making, reasoning, drafting substantive orders, bail or sentencing; AI citations and summaries must be independently verified.", "keyProvisions": [ "AI shall not be used for decision-making, reasoning, drafting of substantive orders, judgment preparation, bail, sentencing or any substantive judicial process (cl. 3.1, 8).", "Permitted: research and precedent identification, translation, administrative support and language refinement of drafts (cl. 7).", "Independent verification of AI-generated citations, statutory references and summaries (cl. 10).", "Public AI tools restricted; sensitive case material must not be uploaded (cl. 9.2-9.3); only High Court-approved tools.", "To be read with the IT Act, the DPDP Act 2023 when in force, the Contempt of Courts Act 1971 and the Gujarat High Court Rules 1993." ], "relatedIncidents": [], "sources": [ { "label": "Gujarat High Court (official PDF)", "url": "https://gujarathighcourt.nic.in/hccms/sites/default/files/miscnotifications/Policy%20on%20the%20use%20of%20Artificial%20Intelligence%20in%20the%20Judicial%20and%20Court%20Administration.pdf" }, { "label": "Verdictum column analysing the policy", "url": "https://www.verdictum.in/columns/policy-on-the-use-of-artificial-intelligence-gujarat-high-court-1613405" } ], "country": "IN", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "court-own-use", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "my-jpdp-admp-guideline", "url": "https://safelegalai.com/regulation/other/my-jpdp-admp-guideline", "title": "Guideline on Automated Decision-Making and Profiling (ADMP) under the Personal Data Protection Act 2010", "body": "Personal Data Protection Department [Jabatan Perlindungan Data Peribadi], Ministry of Digital", "bodyShort": "JPDP", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "supervision", "consent", "competence", "risk-classification" ], "effectiveDate": "2026-04-01T00:00:00.000Z", "versions": [ { "version": "Issued April 2026", "date": "2026-04-01T00:00:00.000Z", "url": "https://www.pdp.gov.my/ppdpv1/wp-content/uploads/2026/04/GARIS-PANDUAN-PEMBUATAN-KEPUTUSAN-SECARA-AUTOMATIK-DAN-PEMPROFILAN-ADMP-1.pdf", "changes": "Only the month of issue (April 2026) is established, so the date is recorded as the first of that month. Issued with companion DPIA and Data Protection by Design guidelines." } ], "summary": "A regulatory guideline under the Personal Data Protection Act 2010 (Act 709) as amended in 2024. Part E governs AI and generative AI in profiling and automated decision-making, requiring human oversight, barring decisions made on a sole automated factor, and requiring staff training and appointed reviewers.", "keyProvisions": [ "Part E applies the PDPA to AI and generative AI used in automated decision-making and profiling.", "Human oversight is required; decisions must not rest on a sole automated factor.", "Data subjects must be informed of AI use in automated decision-making and profiling through the privacy notice.", "Requires measures against over-reliance when developing, supplying or using AI systems and services." ], "relatedIncidents": [], "sources": [ { "label": "JPDP — ADMP guideline (official PDF)", "url": "https://www.pdp.gov.my/ppdpv1/wp-content/uploads/2026/04/GARIS-PANDUAN-PEMBUATAN-KEPUTUSAN-SECARA-AUTOMATIK-DAN-PEMPROFILAN-ADMP-1.pdf" }, { "label": "JPDP — guidelines index", "url": "https://www.pdp.gov.my/ppdpv1/akta/garis-panduan-ms/" } ], "country": "MY", "categories": [ "data-protection", "client-consent-transparency", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "za-ocj-draft-ai-policy-judiciary", "url": "https://safelegalai.com/regulation/south-africa/za-ocj-draft-ai-policy-judiciary", "title": "Draft Policy on Artificial Intelligence for the Judiciary in South Africa", "body": "Office of the Chief Justice / South African Judiciary", "bodyShort": "OCJ", "jurisdiction": "south-africa", "type": "policy", "status": "proposed", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "supervision", "competence", "record-keeping" ], "versions": [ { "version": "Draft for comment", "date": "2026-04-01T00:00:00.000Z", "url": "https://www.judiciary.org.za/images/news/2026/FINAL%20ADOPTED%202026%20Judiciary%20Conference%20Resolutions.pdf", "changes": "Circulated for comment in April 2026; referenced and directed to be finalised by resolution 5.2 of the 2026 Judiciary Conference (16 July 2026). The draft text itself is not published at a stable official URL. Month-only date: the source gives the month, so the 1st is recorded as the date. The draft policy text is not published at a stable official URL; the record rests on the Judiciary Conference resolutions and press, so it is unverified." } ], "summary": "A draft policy circulated for comment by the Office of the Chief Justice in April 2026 setting binding standards for responsible, ethical, secure and transparent AI use in the South African Judiciary. It distinguishes adjudicative from administrative functions, prohibits AI from adjudicative work, and imposes an explainability requirement. Not adopted as of 4 September 2026.", "keyProvisions": [ "AI systems are assistive tools only and may not substitute for judicial reasoning, discretion or decision-making.", "AI is expressly prohibited from any adjudicative function, including evaluating evidence, assessing witness credibility, interpreting or applying law, generating substantive legal reasoning, and determining bail, sentence or risk.", "Any AI system deployed in the Judiciary must produce outputs that are intelligible and open to scrutiny (explainability requirement).", "Draws a clear distinction between adjudicative and administrative functions." ], "relatedIncidents": [], "sources": [ { "label": "South African Judiciary — 2026 Judiciary Conference Resolutions, s5.2 (official PDF, records the draft policy and directs its finalisation)", "url": "https://www.judiciary.org.za/images/news/2026/FINAL%20ADOPTED%202026%20Judiciary%20Conference%20Resolutions.pdf" }, { "label": "IOL — AI in the courtroom: South Africa's new policy strikes a balance (secondary, 8 May 2026)", "url": "https://iol.co.za/news/2026-05-08-ai-in-the-courtroom-south-africas-new-policy-strikes-a-balance/" } ], "country": "ZA", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ccbe-technical-guide-ai-tools-lawyers", "url": "https://safelegalai.com/regulation/eu/ccbe-technical-guide-ai-tools-lawyers", "title": "CCBE Technical guide on the use of AI tools and models by lawyers (Edition 2026)", "body": "Council of Bars and Law Societies of Europe (CCBE)", "bodyShort": "CCBE", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "permissive" ], "effectiveDate": "2026-03-27T00:00:00.000Z", "versions": [ { "version": "Edition 2026", "date": "2026-03-27T00:00:00.000Z", "url": "https://www.ccbe.eu/fileadmin/speciality_distribution/public/documents/IT_LAW/ITL_Guides_recommendations/EN_ITL_20260327_CCBE-technical-guide-on-the-use-of-AI-tools-and-models-by-lawyers.pdf" } ], "summary": "Companion to the CCBE generative-AI guide and cloud-computing guidelines. It explains deployment options - on-premises, self-hosted colocation, bring-your-own-model and fully managed SaaS - and their confidentiality and data-residency implications, model capability, size, speed and context-length considerations, which models can run locally, and a technical glossary grounding the duty to understand the tools used.", "keyProvisions": [ "Lawyers have a professional duty to understand the essential functioning of the technologies they use.", "Compares hosting architectures by confidentiality and control.", "Technical criteria for selecting and evaluating AI tools, with no product endorsements." ], "relatedIncidents": [], "sources": [ { "label": "CCBE — Technical guide (PDF)", "url": "https://www.ccbe.eu/fileadmin/speciality_distribution/public/documents/IT_LAW/ITL_Guides_recommendations/EN_ITL_20260327_CCBE-technical-guide-on-the-use-of-AI-tools-and-models-by-lawyers.pdf" }, { "label": "Croatian Bar Association republication", "url": "https://www.hok-cba.hr/hok/tehnicki-vodic-ccbe-a-o-primjeni-alata-i-modela-umjetne-inteligencije-u-odvjetnistvu/" } ], "country": "EU", "categories": [ "procurement-vendor", "confidentiality-client-data", "competence-duty", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "israel-courts-ai-ethics-code-01-26", "url": "https://safelegalai.com/regulation/israel/israel-courts-ai-ethics-code-01-26", "title": "Ethical Code for the Use of Artificial Intelligence Tools in the Judicial Authority (Courts Administration Procedure No. 01-26)", "body": "Courts Administration [Hanhalat Batei HaMishpat], Judicial Authority of Israel", "bodyShort": "Israel Courts Administration", "jurisdiction": "israel", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "disclosure", "supervision", "competence", "prohibition" ], "versions": [ { "version": "Procedure No. 01-26", "date": "2026-03-26T00:00:00.000Z", "url": "https://journal.lawforum.org.il/gruzman-ai/", "changes": "First Israeli judicial AI instrument. Issued by Courts Administration Director Judge Tzachi Uziel, drafted by a team chaired by Judge Rami Khaimowitz of the Central District Court, approved by Supreme Court President Yitzhak Amit. Sources are secondary (law journal and press) quoting the procedure; the official text was not obtained, so this record is unverified." } ], "summary": "A binding internal code setting principles, operative instructions and control mechanisms for efficient, careful and responsible use of AI tools across the Israeli court system. It covers judges, registrars and administrative staff only, not lawyers or litigants, and was issued after a run of decisions exposing AI-fabricated citations.", "keyProvisions": [ "AI is an aid not intended to replace the human element; it does not diminish the discretion or responsibility of judges and court staff.", "Users must verify every factual determination, quotation, reference to legislation and case law, legal summary and other information against authoritative sources.", "Material from case files managed in Net HaMishpat may not be entered into AI tools that are not internal to the judiciary.", "The court system's use of AI is to be visible to the public, with disclosure rules tailored to each system.", "Use of AI tools is conditional on appropriate prior training.", "Administrative or registry AI must preserve litigants' rights and procedural equality; where an automated action may affect rights, an objection route must be available." ], "relatedIncidents": [], "sources": [ { "label": "Rashut HaRabim law journal, citing Procedure 01-26 and quoting its text", "url": "https://journal.lawforum.org.il/gruzman-ai/" }, { "label": "CTech, Israeli courts AI ethics code (20 Apr 2026)", "url": "https://www.calcalistech.com/ctechnews/article/bj8k2yqp11e" }, { "label": "Bizportal/Takdin, quoting the code", "url": "https://www.bizportal.co.il/takdin/news/article/20029582" }, { "label": "Globes, the drafting task force", "url": "https://www.globes.co.il/news/article.aspx?did=1001516452" } ], "country": "IL", "categories": [ "judicial-use", "court-own-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "disclosure-filings", "supervision", "training-cle", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "fr-cnb-guide-deontologie-ia", "url": "https://safelegalai.com/regulation/other/fr-cnb-guide-deontologie-ia", "title": "Guide pratique — La déontologie et l'intelligence artificielle", "body": "Conseil national des barreaux (National Bar Council), France", "bodyShort": "CNB", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "consent", "permissive" ], "versions": [ { "version": "Adopted at the assemblée générale of 17 March 2026", "date": "2026-03-17T00:00:00.000Z", "url": "https://cnb.avocat.fr/actualite/le-cnb-adopte-un-guide-sur-la-deontologie-et-l-intelligence-artificielle", "changes": "First CNB deontological framework specific to generative AI, following the September 2024 practical guide. The guide itself is not public; the record rests on the CNB's own announcement, so it is unverified." } ], "summary": "The first explicit French deontological framework for generative AI in law firms, adopted by the CNB's general assembly on 17 March 2026. It permits AI on condition that the lawyer keeps intellectual control and responsibility, and applies professional secrecy, GDPR compliance, competence, prudence, independence, client information and balanced fee-setting to AI-assisted work, with recent case-law examples.", "keyProvisions": [ "The lawyer keeps 'la maîtrise intellectuelle' and full responsibility for AI-assisted work in all circumstances.", "Protection of the secret professionnel and GDPR compliance frame what may be entered into a tool.", "Obligations of competence, prudence and independence apply to AI use.", "Clients must be informed, and fees must be set in a balanced way where AI reduces the work involved.", "Includes recent French court decisions on AI hallucinations as worked examples." ], "relatedIncidents": [], "sources": [ { "label": "CNB — Le CNB adopte un guide sur la déontologie et l'Intelligence artificielle (official)", "url": "https://cnb.avocat.fr/actualite/le-cnb-adopte-un-guide-sur-la-deontologie-et-l-intelligence-artificielle" } ], "country": "FR", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "client-consent-transparency", "fees-billing", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "eg-national-ai-governance-framework-guide", "url": "https://safelegalai.com/regulation/other/eg-national-ai-governance-framework-guide", "title": "Guide to Egypt's National AI Governance Framework, Edition 2.0", "body": "Egyptian Center for Responsible AI (ECRAI), under the National Council for Artificial Intelligence, Quantum Computing and Emerging Technologies", "bodyShort": "ECRAI / Egypt NCAI", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "risk-classification", "disclosure", "competence" ], "versions": [ { "version": "2.0", "date": "2026-03-15T00:00:00.000Z", "url": "https://ai.gov.eg/SynchedFiles/ar/Resources/Guide%20to%20Egypt%E2%80%99s%20National%20AI%20Governance%20Framework%2015-3-2026.pdf", "changes": "Approved for publication March 2026; document ID EG-NCAI-GOV-FW-2026-02. Drafted from version 0.1 (Aug 2025) and 0.8 (Oct 2025)." } ], "summary": "Egypt's umbrella AI governance guide, published March 2026 by the Egyptian Center for Responsible AI. It defines a four-tier risk classification with a mandatory dual-check conformity model for high-risk systems, drawing on UNESCO, OECD, ISO/IEC and the G7 Hiroshima Process. It does not name justice or the courts.", "keyProvisions": [ "Tier 1 (prohibited): AI creating unacceptable risk to individuals' safety, livelihoods or fundamental rights.", "Tier 2 (high risk): critical infrastructure and biometrics; pre- and post-market conformity assessment under a mandatory dual-check compliance model before placement on the Egyptian market.", "Tier 3 (limited risk): chatbots, emotion recognition and generative AI including deepfakes; transparency obligations, disclosure of machine interaction and labelling or watermarking of AI-generated content.", "Tier 4 (minimal risk): no additional regulatory obligations; voluntary code of conduct.", "Full-text search of the guide returns no reference to courts, judges, the judiciary, litigation or the legal profession; administration of justice is not listed as a high-risk domain." ], "relatedIncidents": [], "sources": [ { "label": "Official PDF (ai.gov.eg)", "url": "https://ai.gov.eg/SynchedFiles/ar/Resources/Guide%20to%20Egypt%E2%80%99s%20National%20AI%20Governance%20Framework%2015-3-2026.pdf" } ], "country": "EG", "categories": [ "ai-act-horizontal", "evidence-admissibility", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-state-bar-proposed-ai-rule-amendments", "url": "https://safelegalai.com/regulation/us-state/ca-state-bar-proposed-ai-rule-amendments", "title": "Proposed amendments to Rules of Professional Conduct 1.1, 1.4, 1.6, 3.3, 5.1 and 5.3 related to artificial intelligence", "body": "State Bar of California", "bodyShort": "State Bar of California", "jurisdiction": "us-state", "region": "California", "type": "consultation", "status": "proposed", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "supervision", "competence", "confidentiality" ], "versions": [ { "version": "Public comment draft (COPRAC approved 13 March 2026)", "date": "2026-03-13T00:00:00.000Z", "url": "https://www.calbar.ca.gov/public/public-meetings-comment/public-comment/public-comment-archives/2026-public-comment/proposed-amendments-rules-professional-conduct-related-artificial-intelligence" } ], "summary": "Proposed amendments to the California Rules of Professional Conduct that would write AI duties directly into the rules: verification of cited authorities, firm-level AI governance policies under Rule 5.1, and instruction of nonlawyer assistants under Rule 5.3. Public comment closed 4 May 2026 and the Board of Trustees considered them in mid May 2026. Adoption requires the California Supreme Court.", "keyProvisions": [ "Would amend Rule 3.3 to require verification that authorities cited to a tribunal exist and support the proposition.", "Would amend Rule 5.1 to require managerial lawyers to adopt AI governance policies.", "Would amend Rule 5.3 to require instruction of nonlawyer staff on AI use.", "Would add AI competence and confidentiality comments to Rules 1.1, 1.4 and 1.6.", "Not yet adopted by the California Supreme Court as of September 2026." ], "relatedIncidents": [ "noland-v-land-of-the-free" ], "sources": [ { "label": "State Bar of California — proposed AI rule amendments (public comment)", "url": "https://www.calbar.ca.gov/public/public-meetings-comment/public-comment/public-comment-archives/2026-public-comment/proposed-amendments-rules-professional-conduct-related-artificial-intelligence" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "int-lcia-rules-revision-consultation-2026", "url": "https://safelegalai.com/regulation/international/int-lcia-rules-revision-consultation-2026", "title": "LCIA Arbitration Rules revision consultation (AI disclosure and guardrails)", "body": "London Court of International Arbitration", "bodyShort": "LCIA", "jurisdiction": "international", "type": "consultation", "status": "consultation", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "disclosure" ], "versions": [ { "version": "Stage 1 consultation", "date": "2026-03-11T00:00:00.000Z", "url": "https://www.lcia.org/News/lcia-rules-revision-consultation.aspx", "changes": "Submissions closed 11 May 2026; draft revised Arbitration and Mediation Rules due for a second round later in 2026." } ], "summary": "On 11 March 2026 the LCIA opened consultation on the next edition of its 2020 Arbitration Rules. Listed themes include AI disclosure and guardrails alongside digital-ready proceedings, cybersecurity and data protection. Stage 1 closed on 11 May 2026 and draft rules will be published for further comment later in 2026. The LCIA currently has no AI-specific guidance.", "keyProvisions": [ "Consultation theme: AI disclosure and guardrails in LCIA proceedings.", "Consultation theme: digital-ready proceedings, cybersecurity and data protection.", "No AI-specific LCIA rule or guidance is in force pending the revision." ], "relatedIncidents": [], "sources": [ { "label": "LCIA — Rules Revision Consultation", "url": "https://www.lcia.org/News/lcia-rules-revision-consultation.aspx" } ], "country": "INT", "categories": [ "arbitration-adr", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ke-hc-savla-ai-pleadings-rulings", "url": "https://safelegalai.com/regulation/other/ke-hc-savla-ai-pleadings-rulings", "title": "High Court of Kenya rulings on AI-generated pleadings: Savla v Commission on Administrative Justice and the April 2026 set-aside", "body": "High Court of Kenya at Milimani", "bodyShort": "High Court of Kenya", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "litigants", "lawyers", "parties" ], "requirements": [ "verification", "prohibition" ], "versions": [ { "version": "Ruling striking out AI-generated application (Mwamuye J)", "date": "2026-03-06T00:00:00.000Z", "url": "https://capitalfm.africa/milimani-high-court-ai-generated-documents-struck-out/", "changes": "Notice of motion and supporting affidavit found machine-generated and struck out for breach of Order 51 rule 13 and Order 19 rules 4-5 of the Civil Procedure Rules; defects held incurable by amendment, fresh compliant filing permitted. The Kenya Law page for the April 2026 ruling returned 404 on 4 September 2026 and the account rests partly on press reports, so this record is unverified." }, { "version": "Ruling setting aside the 23 December 2025 judgment (Chigiti J)", "date": "2026-04-16T00:00:00.000Z", "url": "https://new.kenyalaw.org/akn/ke/judgment/kehc/2026/5485/eng@2026-04-16", "changes": "Judgment of 23 December 2025 set aside; the AI-founded application struck out with costs against the applicant." } ], "summary": "Two 2026 High Court rulings in the same litigation set Kenya's de facto position: AI-generated pleadings are procedurally deficient. Mwamuye J held that computer-generated documents or AI outputs cannot substitute for human-drawn documents; Chigiti J then set aside the resulting judgment, reasoning that the Civil Procedure Rules do not authorise AI-drafted pleadings and that their use confers an unfair advantage.", "keyProvisions": [ "Mwamuye J: computer-generated documents or outputs of artificial intelligence cannot be a proper substitute for human-drawn documents.", "The offending notice of motion breached Order 51 rule 13 and the supporting affidavit Order 19 rules 4 and 5 of the Civil Procedure Rules 2010; the defects could not be cured by amendment.", "Chigiti J: drafting of pleadings is regulated by Order 2 of the Civil Procedure Rules, which does not authorise AI-generated content.", "Chigiti J: use of AI confers an unfair advantage on one party and undermines the adversarial system; the integrity of the process is compromised when unknown tools influence judicial outcomes (Article 48, Constitution of Kenya).", "Orders: the 23 December 2025 judgment set aside and the application struck out with costs against the applicant." ], "relatedIncidents": [], "sources": [ { "label": "Commission on Administrative Justice v Savla [2026] KEHC 5485 (KLR) (Kenya Law index entry; the AKN page returned 404 on 4 Sep 2026)", "url": "https://new.kenyalaw.org/akn/ke/judgment/kehc/2026/5485/eng@2026-04-16" }, { "label": "Savla v Commission on Administrative Justice [2025] KEHC 19318 (KLR), the judgment later set aside (Kenya Law, opened and confirmed)", "url": "https://new.kenyalaw.org/akn/ke/judgment/kehc/2025/19318/eng@2025-12-23" }, { "label": "Milimani court strikes out AI-generated application (Capital FM, 11 Mar 2026)", "url": "https://capitalfm.africa/milimani-high-court-ai-generated-documents-struck-out/" }, { "label": "High Court declares AI-drafted pleadings illegal (HapaKenya, 19 Apr 2026)", "url": "https://hapakenya.com/2026/04/19/high-court-declares-that-legal-pleadings-drafted-by-ai-are-illegal-in-kenya/" } ], "country": "KE", "categories": [ "verification-duty", "litigant-in-person", "sanctions-record", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "sg-minlaw-generative-ai-guide-legal-sector", "url": "https://safelegalai.com/regulation/singapore/sg-minlaw-generative-ai-guide-legal-sector", "title": "Guide for Using Generative AI in the Legal Sector", "body": "Ministry of Law, Singapore", "bodyShort": "MinLaw", "jurisdiction": "singapore", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "confidentiality", "verification", "disclosure", "competence" ], "effectiveDate": "2026-03-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-03-06T00:00:00.000Z", "url": "https://www.mlaw.gov.sg/files/Guide_for_using_Generative_AI_in_the_Legal_Sector__Published_on_6_Mar_2026_.pdf" } ], "summary": "A guide published by Singapore's Ministry of Law on 6 March 2026 for law practices adopting generative AI. It sets principles of professional ethics, confidentiality and transparency, calls for human-in-the-loop review of high-risk tasks such as court filings and advice, and encourages firms to adopt governance policies and staff training before deployment.", "keyProvisions": [ "Human review is required for high-risk uses such as court documents and legal advice.", "Client confidentiality and data protection obligations apply to information entered into tools.", "Firms should be transparent with clients about material AI use.", "Practices should adopt AI governance policies and train staff." ], "relatedIncidents": [], "article": "singapore-ai-guidance-history", "sources": [ { "label": "Ministry of Law Singapore — Guide for using generative AI in the legal sector (PDF)", "url": "https://www.mlaw.gov.sg/files/Guide_for_using_Generative_AI_in_the_Legal_Sector__Published_on_6_Mar_2026_.pdf" } ], "country": "SG", "categories": [ "bar-guidance", "competence-duty", "confidentiality-client-data", "client-consent-transparency", "supervision", "procurement-vendor", "evidence-admissibility", "training-cle" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "tr-tbmm-ai-commission-report-260", "url": "https://safelegalai.com/regulation/other/tr-tbmm-ai-commission-report-260", "title": "Report of the Parliamentary Research Commission on Artificial Intelligence, Report No. 260", "body": "Grand National Assembly of Türkiye, AI Research Commission", "bodyShort": "TBMM AI Research Commission", "jurisdiction": "other", "type": "report", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Report No. 260, 28th legislative term", "date": "2026-03-01T00:00:00.000Z", "url": "https://www.tmmob.org.tr/sites/default/files/tbmm_yapay_zeka_arastirma_komisyon_raporu.pdf", "changes": "First institutional parliamentary document setting out Türkiye's AI law debate comprehensively. Exact day of publication within March 2026 not established. Month-only date: the source gives the month, so the 1st is recorded as the date. The official TBMM copy was not reachable; the record rests on a mirrored PDF and a secondary legal assessment, so it is unverified." } ], "summary": "The parliamentary AI research commission's March 2026 report, the first comprehensive institutional statement of Türkiye's AI regulation debate. It recommends a framework Turkish AI law, a Turkish AI Authority, regulatory sandboxes, a text-and-data-mining exception in copyright law, modernisation of the data protection regime, and ratification of the Council of Europe AI Framework Convention.", "keyProvisions": [ "Recommends a framework Turkish artificial intelligence law.", "Recommends establishing a Turkish AI Authority as sector regulator.", "Recommends regulatory sandboxes for AI experimentation.", "Recommends a data-mining exception in copyright law (FSEK) and modernisation of the personal data protection framework.", "Recommends ratification of the Council of Europe Framework Convention on AI." ], "relatedIncidents": [], "sources": [ { "label": "SETA legal assessment of the report (June 2026)", "url": "https://www.setav.org/turk-yapay-zeka-kanununa-dogru-tbmm-raporunun-hukuki-degerlendirmesi" }, { "label": "Report PDF (copy hosted by TMMOB)", "url": "https://www.tmmob.org.tr/sites/default/files/tbmm_yapay_zeka_arastirma_komisyon_raporu.pdf" } ], "country": "TR", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "es-tsj-canarias-2026-48-citas-falsas", "url": "https://safelegalai.com/regulation/other/es-tsj-canarias-2026-48-citas-falsas", "title": "TSJ Canarias, Criminal Chamber (Feb 2026): fine for 48 AI-fabricated Supreme Court citations", "body": "High Court of Justice of the Canary Islands (Tribunal Superior de Justicia de Canarias)", "bodyShort": "TSJ Canarias", "jurisdiction": "other", "region": "Canary Islands", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "February 2026", "date": "2026-02-28T00:00:00.000Z", "url": "https://www.xataka.com/magnet/tribunal-superior-justicia-canarias-multa-a-abogado-citar-48-sentencias-falsas-se-habia-sugerido-ia", "changes": "Reported in press; exact date and official text not obtained." } ], "summary": "First effective Spanish sanction for AI hallucinations. On appeal from the Audiencia Provincial de Santa Cruz de Tenerife, counsel cited 48 Supreme Court rulings and a CGPJ report that did not exist, sourced from ChatGPT without checking CENDOJ. The chamber fined the lawyer about EUR 420-450 for procedural bad faith under Article 247 LEC.", "keyProvisions": [ "Fabricated citations constitute procedural bad faith under Art. 247 LEC.", "Fine pegged to roughly half the annual cost of a reliable legal-AI subscription.", "Reported referral to the bar for disciplinary review." ], "relatedIncidents": [], "sources": [ { "label": "Xataka (secondary)", "url": "https://www.xataka.com/magnet/tribunal-superior-justicia-canarias-multa-a-abogado-citar-48-sentencias-falsas-se-habia-sugerido-ia" }, { "label": "Noticias Juridicas (secondary)", "url": "https://noticias.juridicas.com/actualidad/noticias/20771-el-tsj-canarias-investiga-a-un-abogado-por-citar-sentencias-falsas-creadas-con-ia/" } ], "country": "ES", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "kr-scourt-ai-guidebook-for-judges-2026", "url": "https://safelegalai.com/regulation/other/kr-scourt-ai-guidebook-for-judges-2026", "title": "AI Guidebook for Judges", "body": "National Court Administration, Supreme Court of Korea", "bodyShort": "Korea National Court Administration", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "verification", "confidentiality", "competence" ], "effectiveDate": "2026-02-24T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-02-24T00:00:00.000Z", "url": "https://dailian.co.kr/news/view/1613520/", "changes": "Announced 24 February 2026; distributed to all judges in March 2026. Official press release not retrievable." } ], "summary": "Guidebook prepared by a National Court Administration task force of eight judges and distributed to all Korean judges in March 2026. It sets standards and about twenty graded worked examples for using commercial AI in trial and administrative work, with checklists on hallucination, data bias, personal-data breaches and security, and prompt-writing guidance.", "keyProvisions": [ "Judges may use commercial AI for document management, drafting support, research and dispute analysis within stated limits.", "Checklists for hallucination, bias, privacy and security risks.", "Judicial responsibility and impartiality preserved; AI cannot replace judicial judgment." ], "relatedIncidents": [], "sources": [ { "label": "Dailian report", "url": "https://dailian.co.kr/news/view/1613520/" }, { "label": "Newspim on the judiciary AI platform (7 August 2026)", "url": "https://www.newspim.com/news/view/20260806001025" } ], "country": "KR", "categories": [ "judicial-use", "verification-duty", "ai-decision-prohibition", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "nzls-generative-ai-guidance-lawyers", "url": "https://safelegalai.com/regulation/nz/nzls-generative-ai-guidance-lawyers", "title": "Generative AI guidance for lawyers", "body": "New Zealand Law Society", "bodyShort": "NZ Law Society", "jurisdiction": "nz", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence" ], "effectiveDate": "2024-01-01T00:00:00.000Z", "versions": [ { "version": "February 2026 update", "date": "2026-02-24T00:00:00.000Z", "url": "https://www.lawsociety.org.nz/professional-practice/rules-and-maintaining-professional-standards/generative-ai-guidance-for-lawyers/", "changes": "Updated guidance on verification, privacy and privilege, client communication and fee review; first published in 2024." }, { "version": "Original (2024)", "date": "2024-01-01T00:00:00.000Z", "url": "https://www.lawsociety.org.nz/professional-practice/rules-and-maintaining-professional-standards/generative-ai-guidance-for-lawyers/" } ], "summary": "Guidance from the New Zealand Law Society, updated 24 February 2026, on how the Lawyers and Conveyancers Act rules apply to generative AI. Lawyers must verify all output, protect privacy and privilege when entering information, disclose AI use to clients where appropriate, and review whether fees remain reasonable when AI reduces the time spent.", "keyProvisions": [ "All AI-generated content must be verified before reliance.", "Privacy and legal privilege must be protected in any data entered into tools.", "Clients should be told about AI use where appropriate to the engagement.", "Fees should be reviewed where AI materially reduces time spent." ], "relatedIncidents": [ "jones-v-family-court-whangarei" ], "sources": [ { "label": "New Zealand Law Society — generative AI guidance for lawyers", "url": "https://www.lawsociety.org.nz/professional-practice/rules-and-maintaining-professional-standards/generative-ai-guidance-for-lawyers/" } ], "country": "NZ", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "client-consent-transparency", "fees-billing", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "it-tribunale-siracusa-338-2026-ai-hallucinations", "url": "https://safelegalai.com/regulation/other/it-tribunale-siracusa-338-2026-ai-hallucinations", "title": "Tribunale di Siracusa, Sez. II civ., judgment no. 338/2026 (AI-hallucinated citations, Art. 96 c.p.c.)", "body": "Court of Siracusa (Tribunale di Siracusa)", "bodyShort": "Trib. Siracusa", "jurisdiction": "other", "region": "Sicily", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "Judgment 20 February 2026", "date": "2026-02-20T00:00:00.000Z", "url": "https://ntplusdiritto.ilsole24ore.com/art/quattro-sentenze-fantasma-e-conto-30000-euro-quando-l-utilizzazione-acritica-dell-ia-entra-tribunale-AIW7MlmB", "changes": "Reported by Il Sole 24 Ore NT+ Diritto; official text not obtained." } ], "summary": "In a contract dispute counsel cited four Cassazione decisions whose numbers or content did not exist, apparently generated by AI without checking the CED database. The court held that uncritical use of generative AI is gross negligence rather than a mitigating factor and applied Article 96(3)-(4) c.p.c., ordering EUR 30,200 in fees, damages and a court-fund payment.", "keyProvisions": [ "Unverified generative-AI output in a filing is gross negligence under Art. 96(3) c.p.c.", "Damages for reckless litigation plus a payment to the Cassa delle ammende under Art. 96(4).", "Contrasts with Trib. Firenze (14 March 2025), which refused Art. 96 sanctions absent bad faith." ], "relatedIncidents": [], "sources": [ { "label": "Il Sole 24 Ore NT+ Diritto (secondary)", "url": "https://ntplusdiritto.ilsole24ore.com/art/quattro-sentenze-fantasma-e-conto-30000-euro-quando-l-utilizzazione-acritica-dell-ia-entra-tribunale-AIW7MlmB" }, { "label": "Rivista AI survey of Italian hallucination cases (secondary)", "url": "https://www.rivista.ai/2026/08/26/ai-e-precedenti-giurisprudenziali-inventati-come-la-giurisprudenza-italiana-sta-ridefinendo-la-responsabilita-degli-avvocati/" } ], "country": "IT", "categories": [ "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ke-artificial-intelligence-bill-2026", "url": "https://safelegalai.com/regulation/other/ke-artificial-intelligence-bill-2026", "title": "The Artificial Intelligence Bill, 2026 (Senate Bills No. 4 of 2025)", "body": "Senate of Kenya", "bodyShort": "Kenyan Senate", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification", "record-keeping", "disclosure", "consent", "prohibition" ], "versions": [ { "version": "As published", "date": "2026-02-19T00:00:00.000Z", "url": "https://new.kenyalaw.org/akn/ke/bill/senate/2026-02-19/the-artificial-intelligence-bill-2026/eng@2026-02-19", "changes": "Published as a Senate Bill; must also be considered by the National Assembly because it concerns county governments." } ], "summary": "Kenya's horizontal AI bill, published 19 February 2026. It creates an Office of the Artificial Intelligence Commissioner, a four-tier risk classification set by regulations, obligations for high-risk systems including human oversight and human rights impact assessments, a public register of high-risk systems, and regulatory sandboxes. Justice and court uses are not named.", "keyProvisions": [ "Classification: unacceptable risk (prohibited), high risk for critical sectors including healthcare, education, agriculture, finance, security, employment or public administration, limited risk and minimal risk; categories to be prescribed by the Cabinet Secretary on the Commissioner's recommendation.", "Clause 26: high-risk providers and deployers must conduct a pre-deployment risk assessment with human oversight, a human rights impact assessment, ensure transparency, traceability and explainability, and keep records of inputs, training datasets and outputs for at least five years.", "Clause 26(1)(g): systems generating or manipulating images, voice or likeness require explicit consent and clear AI-generated labelling.", "Clause 28: automated decisions with significant legal effects must carry Data Protection Act safeguards including rights to human intervention, to express views and to contest; annual compliance reports go to the Commissioner.", "Clause 27: a public register of high-risk AI systems, including those used by county governments." ], "relatedIncidents": [], "sources": [ { "label": "The Artificial Intelligence Bill, 2026 (Kenya Law official)", "url": "https://new.kenyalaw.org/akn/ke/bill/senate/2026-02-19/the-artificial-intelligence-bill-2026/eng@2026-02-19" } ], "country": "KE", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "es-aepd-orientaciones-ia-agentica-2026", "url": "https://safelegalai.com/regulation/other/es-aepd-orientaciones-ia-agentica-2026", "title": "AEPD Guidance on agentic AI from a data-protection perspective (Orientaciones sobre IA agentica)", "body": "Spanish Data Protection Agency (Agencia Espanola de Proteccion de Datos)", "bodyShort": "AEPD", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "risk-classification", "supervision", "confidentiality" ], "effectiveDate": "2026-02-18T00:00:00.000Z", "versions": [ { "version": "18 February 2026", "date": "2026-02-18T00:00:00.000Z", "url": "https://www.aepd.es/guias/orientaciones-ia-agentica.pdf" } ], "summary": "Guidance for controllers and processors deploying AI agents that plan and execute multi-step tasks with personal data. It identifies agent-specific risks such as autonomy, tool access and chained actions, and directs risk assessment and DPIAs, human oversight, data minimisation and clear processor arrangements. Not legal-sector specific but relied on by Spanish bars for AI use in firms.", "keyProvisions": [ "Agentic systems require risk analysis beyond prompt-level supervision.", "Human oversight and limits on autonomous actions.", "Data minimisation and purpose limitation across tool calls.", "Contractual allocation of controller and processor roles with providers." ], "relatedIncidents": [], "sources": [ { "label": "AEPD — Orientaciones sobre IA agentica (PDF)", "url": "https://www.aepd.es/guias/orientaciones-ia-agentica.pdf" }, { "label": "ICAM summary", "url": "https://web.icam.es/orientaciones-de-la-aepd-sobre-inteligencia-artificial-agentica-desde-la-perspectiva-de-proteccion-de-datos/" } ], "country": "ES", "categories": [ "data-protection", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ph-sc-am-25-11-28-ai-governance-framework", "url": "https://safelegalai.com/regulation/other/ph-sc-am-25-11-28-ai-governance-framework", "title": "A.M. No. 25-11-28-SC: Governance Framework on the Use of Human-Centered Augmented Intelligence in the Judiciary", "body": "Supreme Court of the Philippines (En Banc)", "bodyShort": "Supreme Court of the Philippines", "jurisdiction": "other", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges", "court-staff", "lawyers", "litigants", "providers", "all" ], "requirements": [ "disclosure", "verification", "prohibition", "confidentiality", "supervision", "record-keeping", "competence" ], "effectiveDate": "2026-02-18T00:00:00.000Z", "versions": [ { "version": "Adopted by the En Banc", "date": "2026-02-18T00:00:00.000Z", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/101138" } ], "summary": "The Supreme Court of the Philippines' judiciary-wide AI framework, adopted 18 February 2026 and announced on 19 March 2026. It binds justices and judges, court personnel, court users including members of the Bar and the public, and AI vendors, and creates a permanent Committee on Human-Centered Augmented Intelligence.", "keyProvisions": [ "AI tools or their output must under no circumstance serve as the sole, primary or determinative basis of any adjudicatory outcome.", "Mandatory AI-use disclosure for court-issued adjudicatory documents: tool and version, purpose, extent of use, degree of human oversight, output preserved, and a responsibility statement.", "No AI tool shall be used to process confidential, privileged, sensitive or secret information without express authority and under strict supervision of the Court.", "No AI tool may be used in the judiciary unless authorised by the Supreme Court En Banc, following a phased pilot-then-rollout process.", "A lawyer shall not evade responsibility by claiming that a breach was the fault of, or caused by, an AI tool.", "Developers must disclose a tool's logic, limitations and safeguards; trade-secret claims limiting disclosure must be justified; the AI Committee may require ethical impact assessments, adversarial red-teaming and bug-bounty testing.", "PHILJA is directed to run AI certificate training for the Bench; the IBP is directed to monitor the profession's AI use and report to the Court." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court E-Library — A.M. No. 25-11-28-SC full text", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/101138" }, { "label": "Supreme Court — SC adopts framework for the judiciary's use of AI (19 Mar 2026)", "url": "https://sc.judiciary.gov.ph/sc-adopts-framework-for-judiciarys-use-of-artificial-intelligence/" } ], "country": "PH", "categories": [ "disclosure-filings", "verification-duty", "judicial-use", "ai-decision-prohibition", "litigant-in-person", "confidentiality-client-data", "supervision", "court-own-use", "procurement-vendor", "training-cle", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ok-occa-rule-1-17-generative-ai", "url": "https://safelegalai.com/regulation/us-state/us-ok-occa-rule-1-17-generative-ai", "title": "Oklahoma Court of Criminal Appeals Rule 1.17 — Use of Generative Artificial Intelligence", "body": "Oklahoma Court of Criminal Appeals", "bodyShort": "Oklahoma Court of Criminal Appeals", "jurisdiction": "us-state", "region": "Oklahoma", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "certification" ], "versions": [ { "version": "Original", "date": "2026-02-18T00:00:00.000Z", "url": "https://www.okcca.net/cases/2026/OK-CR-7/", "changes": "Adopted by order In re: Addition of a New Rule to the Rules of the Oklahoma Court of Criminal Appeals, 2026 OK CR 7" } ], "summary": "Rule 1.17 requires that where generative AI has been used in drafting any document filed in the Oklahoma Court of Criminal Appeals, the party or counsel must ensure every AI-produced or AI-modified portion has been verified as accurate by a person responsible for the document. Non-compliance may lead to waiver, striking of the document, or contempt.", "keyProvisions": [ "Applies to any document filed in the Oklahoma Court of Criminal Appeals", "Where generative AI has been used, every portion produced or modified by it must be verified as accurate by a person responsible for the document", "Generative AI is defined as any type of AI that generates content or data in response to a prompt or query by a user", "Failure to comply may result in waiver of the affected issue(s) on appeal, striking of the non-compliant document from the record, and/or a finding of contempt" ], "relatedIncidents": [], "sources": [ { "label": "2026 OK CR 7 (official)", "url": "https://www.okcca.net/cases/2026/OK-CR-7/" }, { "label": "OCCA Court Rules index (official)", "url": "https://www.okcca.net/rules/" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "uk-cjc-ai-court-documents-consultation", "url": "https://safelegalai.com/regulation/uk/uk-cjc-ai-court-documents-consultation", "title": "Consultation: The Use of AI for the Preparation of Court Documents", "body": "Civil Justice Council", "bodyShort": "CJC", "jurisdiction": "uk", "type": "consultation", "status": "consultation", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "verification" ], "effectiveDate": "2026-02-17T00:00:00.000Z", "versions": [ { "version": "Consultation paper", "date": "2026-02-17T00:00:00.000Z", "url": "https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/latest-news/" } ], "summary": "The Civil Justice Council consulted from 17 February to 14 April 2026 on whether the Civil Procedure Rules for England and Wales should impose AI-specific requirements on court documents, including transparency or disclosure duties for pleadings, skeleton arguments, witness statements and expert reports. Interim findings were published on 30 June 2026; the final report is pending.", "keyProvisions": [ "Asked whether the CPR should require parties to disclose AI use in preparing court documents.", "Covered professional drafting, witness statements, expert evidence and litigants in person separately.", "Consultation closed 14 April 2026.", "Followed by an interim update on findings dated 30 June 2026." ], "relatedIncidents": [ "ayinde-v-haringey" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "Civil Justice Council — latest news (consultation and update)", "url": "https://www.judiciary.uk/related-offices-and-bodies/advisory-bodies/cjc/latest-news/" } ], "country": "GB", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-in-supreme-court-trial-court-ai-toolkit", "url": "https://safelegalai.com/regulation/us-state/us-in-supreme-court-trial-court-ai-toolkit", "title": "Indiana Supreme Court Office of Judicial Administration: trial-court AI policy toolkit", "body": "Indiana Supreme Court, Office of Judicial Administration", "bodyShort": "Indiana Supreme Court OJA", "jurisdiction": "us-state", "region": "Indiana", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "risk-classification", "supervision" ], "versions": [ { "version": "Toolkit", "date": "2026-02-17T00:00:00.000Z", "url": "https://times.courts.in.gov/2026/02/17/ai-toolkit/", "changes": "Released to Indiana trial courts" } ], "summary": "Indiana's judiciary is governing AI administratively rather than by rule. An AI Governance Committee was formed in autumn 2024 and an AI policy toolkit for trial courts was released in February 2026; attorney-facing guidance is still pending. Contents were not verified in this run.", "keyProvisions": [ "Toolkit for Indiana trial courts adopting local AI policies (contents not yet verified)", "Follows creation of the Indiana Supreme Court AI Governance Committee in autumn 2024", "Attorney-facing guidance still pending as of September 2026" ], "relatedIncidents": [], "sources": [ { "label": "Indiana Court Times, AI toolkit", "url": "https://times.courts.in.gov/2026/02/17/ai-toolkit/" }, { "label": "Indiana Court Times, Guardrails for the Future", "url": "https://times.courts.in.gov/2025/08/06/guardrails-for-the-future/" } ], "country": "US", "categories": [ "judicial-use", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ca-on-scj-practice-directions-ai", "url": "https://safelegalai.com/regulation/canada/ca-on-scj-practice-directions-ai", "title": "The Use of Artificial Intelligence (AI) for Court Proceedings (consolidated civil, family and criminal practice directions)", "body": "Ontario Superior Court of Justice", "bodyShort": "Ontario Superior Court", "jurisdiction": "canada", "region": "Ontario", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "certification", "record-keeping" ], "versions": [ { "version": "Consolidated practice directions, AI section added", "date": "2026-02-12T00:00:00.000Z", "url": "https://www.ontariocourts.ca/scj/news/new-practice-directions-on-the-responsible-use-of-artificial-intelligence-in-court-proceedings/", "changes": "AI section added to the consolidated civil, family and criminal provincial practice directions; civil direction last updated 17 March 2026." } ], "summary": "Announced 12 February 2026 and added to the Ontario Superior Court's consolidated civil, family and criminal practice directions. The directions require all legal information obtained with AI assistance to be verified against trusted and authoritative sources, require authorities to be hyperlinked to publicly accessible sites such as CanLII, and set out the consequences of misuse.", "keyProvisions": [ "AI-generated references may include incorrect or fictitious authorities; all AI-assisted legal information must be verified against trusted and authoritative sources.", "Authorities must be hyperlinked to publicly accessible sites such as CanLII, with paragraph-level references.", "Counsel must comply with the factum authenticity certification in r. 4.06.1(2.1) of the Rules of Civil Procedure.", "Misuse may result in cost orders, dismissal of a hearing, contempt proceedings, public reprimand or referral to the Law Society of Ontario." ], "relatedIncidents": [], "sources": [ { "label": "Ontario Superior Court — announcement of new AI practice directions", "url": "https://www.ontariocourts.ca/scj/news/new-practice-directions-on-the-responsible-use-of-artificial-intelligence-in-court-proceedings/" }, { "label": "Ontario Superior Court — consolidated civil provincial practice direction", "url": "https://www.ontariocourts.ca/scj/filing-procedures/provincial/consolidated-civil-provincial-practice-direction/" } ], "country": "CA", "categories": [ "verification-duty", "disclosure-filings", "litigant-in-person", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pt-relacao-guimaraes-2026-acordaos-inexistentes", "url": "https://safelegalai.com/regulation/other/pt-relacao-guimaraes-2026-acordaos-inexistentes", "title": "Tribunal da Relacao de Guimaraes, 10 Feb 2026: six non-existent appellate rulings generated by AI", "body": "Guimaraes Court of Appeal (Tribunal da Relacao de Guimaraes)", "bodyShort": "TRG", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "Acordao 10 February 2026", "date": "2026-02-10T00:00:00.000Z", "url": "https://cnnportugal.iol.pt/inteligencia-artificial/tribunal/relacao-de-guimaraes-acusa-advogado-de-citar-acordaos-que-so-existem-na-imaginacao-da-ia/20260303/69a6c222d34e28842c814a7c", "changes": "Reported by CNN Portugal on 3 March 2026; official text not obtained." } ], "summary": "In a criminal appeal concerning illegal weapons possession the defence cited excerpts from six Relacao and Supreme Court rulings that do not exist. The court called this absolutely inadmissible, explained AI hallucination, held legal professionals to a higher standard, ordered 5 UC in costs and suggested reporting the lawyer to the Ordem dos Advogados.", "keyProvisions": [ "Citing non-existent case law generated by AI is inadmissible and breaches professional duties.", "Costs of 5 UC, about EUR 500, imposed on the appellant.", "Referral to the Portuguese Bar Association suggested." ], "relatedIncidents": [], "sources": [ { "label": "CNN Portugal (secondary)", "url": "https://cnnportugal.iol.pt/inteligencia-artificial/tribunal/relacao-de-guimaraes-acusa-advogado-de-citar-acordaos-que-so-existem-na-imaginacao-da-ia/20260303/69a6c222d34e28842c814a7c" }, { "label": "Publico (secondary)", "url": "https://www.publico.pt/2026/03/03/sociedade/noticia/advogado-acusado-citar-acordaos-inexistentes-ajuda-inteligencia-artificial-2166651" } ], "country": "PT", "categories": [ "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "sra-compliance-tips-ai", "url": "https://safelegalai.com/regulation/uk/sra-compliance-tips-ai", "title": "Compliance tips for solicitors on the use of AI and technology", "body": "Solicitors Regulation Authority", "bodyShort": "SRA", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "supervision", "confidentiality", "verification" ], "effectiveDate": "2022-02-15T00:00:00.000Z", "versions": [ { "version": "Feb 2026 revision", "date": "2026-02-09T00:00:00.000Z", "url": "https://www.sra.org.uk/solicitors/resources/innovate/compliance-tips-for-solicitors/", "changes": "Expanded generative-AI section: explicit expectation that output is checked before reliance, client-confidentiality warnings for public tools, and supervision points for junior staff." }, { "version": "Original", "date": "2022-02-15T00:00:00.000Z", "url": "https://www.sra.org.uk/solicitors/resources/innovate/compliance-tips-for-solicitors/" } ], "summary": "The SRA's compliance tips set out how the Standards and Regulations apply when solicitors use AI and other technology: competence, supervision, confidentiality and verification of output remain the firm's responsibility. Revised 9 February 2026 to address generative AI directly, the page is guidance, not a new rule.", "keyProvisions": [ "Firms remain responsible for work product regardless of the tool used to produce it.", "Output from generative AI must be checked before it is relied on or presented to a court.", "Client information must not be entered into tools that do not offer adequate confidentiality protections.", "Supervisors must understand how staff are using AI and set appropriate controls." ], "relatedIncidents": [ "cork-v-smith", "ayinde-v-haringey", "bandla-v-sra" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "SRA — Compliance tips for solicitors (official page)", "url": "https://www.sra.org.uk/solicitors/resources/innovate/compliance-tips-for-solicitors/" } ], "country": "GB", "categories": [ "bar-guidance", "competence-duty", "supervision", "confidentiality-client-data", "verification-duty", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "int-iba-anlbs-ai-professional-tool-guide", "url": "https://safelegalai.com/regulation/international/int-iba-anlbs-ai-professional-tool-guide", "title": "Guidelines and Regulations to Provide Insights on Public Policies to Ensure AI's Beneficial Use as a Professional Tool", "body": "International Bar Association, Alternative and New Law Business Structures Committee AI Working Group", "bodyShort": "IBA ANLBS", "jurisdiction": "international", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive" ], "versions": [ { "version": "February 2026 update", "date": "2026-02-01T00:00:00.000Z", "url": "https://www.ibanet.org/medias/anlbs-ai-working-group-report-feb-2026-2.pdf", "changes": "Updated country chapters; biennial updates intended." }, { "version": "First edition", "date": "2024-09-18T00:00:00.000Z", "url": "https://www.ibanet.org/PPID/Constituent/Multi-displry_Pract/anlbs-ai-report" } ], "summary": "An IBA committee compendium first published in September 2024 and updated in February 2026, surveying guidelines, ethics opinions and statutes on lawyers' professional use of AI across multilateral organisations and more than 36 jurisdictions. It is descriptive rather than prescriptive but is the broadest cross-border map of bar-level AI guidance.", "keyProvisions": [ "Country-by-country summaries of professional-ethics and liability rules touching AI use by lawyers.", "Notes that general professional-ethics and liability rules apply to AI even without AI-specific rules.", "Includes a chapter on multilateral organisations (UNESCO, OECD, Council of Europe, EU)." ], "relatedIncidents": [], "sources": [ { "label": "IBA — ANLBS AI report page", "url": "https://www.ibanet.org/PPID/Constituent/Multi-displry_Pract/anlbs-ai-report" } ], "country": "INT", "categories": [ "bar-guidance", "unauthorised-practice" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "es-cgae-libro-blanco-ia-abogacia-2026", "url": "https://safelegalai.com/regulation/other/es-cgae-libro-blanco-ia-abogacia-2026", "title": "CGAE White Paper on Artificial Intelligence and the Legal Profession (Libro Blanco sobre IA y Abogacia)", "body": "General Council of Spanish Lawyers (Consejo General de la Abogacia Espanola)", "bodyShort": "CGAE", "jurisdiction": "other", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "supervision", "confidentiality" ], "versions": [ { "version": "Presented 30 January 2026", "date": "2026-01-30T00:00:00.000Z", "url": "https://academiadeia.es/wp-content/uploads/2026/01/libro-ai-digital_compressed-1.pdf", "changes": "Original; mirror copy, official abogacia.es URL not obtained." } ], "summary": "CGAE white paper finding that about 60 per cent of Spanish lawyers already use AI, mapping practical uses and risks and recommending mandatory digital-competence training, informed human supervision and transparency for every AI tool, and protection of professional ethics and client confidentiality. It is non-binding and precedes the binding Circular 3/2026.", "keyProvisions": [ "Every AI tool must be subject to informed human supervision.", "Recommends mandatory digital-competence training for lawyers.", "Frames unsupervised AI use as a source of professional liability." ], "relatedIncidents": [], "sources": [ { "label": "Academia de IA summary and mirror (secondary)", "url": "https://academiadeia.es/libro-blanco-ia-abogacia-descarga/" } ], "country": "ES", "categories": [ "bar-guidance", "training-cle", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "es-cgpj-instruccion-2-2026-ia-actividad-jurisdiccional", "url": "https://safelegalai.com/regulation/other/es-cgpj-instruccion-2-2026-ia-actividad-jurisdiccional", "title": "CGPJ Instruction 2/2026 on the use of AI systems in judicial activity (Instruccion 2/2026)", "body": "General Council of the Judiciary (Consejo General del Poder Judicial)", "bodyShort": "CGPJ", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "confidentiality", "supervision", "permissive" ], "effectiveDate": "2026-01-30T00:00:00.000Z", "versions": [ { "version": "Pleno agreement 28 January 2026, BOE no. 27 of 30 January 2026 (BOE-A-2026-2205)", "date": "2026-01-30T00:00:00.000Z", "url": "https://www.boe.es/diario_boe/txt.php?id=BOE-A-2026-2205", "changes": "Original; in force on publication." } ], "summary": "Binding instruction to Spanish judges and magistrates on AI in jurisdictional work: effective human control, no autonomous decision-making or substitution of judicial reasoning, no AI content without personal critical validation, no unapproved tools except study with open sources and no judicial data, full judicial responsibility, CGPJ supervision of personal-data processing and CGPJ-provided training.", "keyProvisions": [ "Art. 4 principles: effective human control, non-substitution, judicial responsibility, independence, fundamental rights, confidentiality and security, bias prevention, proportionality, training.", "Art. 8 prohibited uses: autonomous decision-making, replacing assessment of facts, evidence or law, incorporating AI content without personal validation, processing specially protected data or profiling beyond legal authorisation.", "Art. 9: only systems supplied by the competent administrations; open tools only for study with open-source information.", "Arts 12-13: CGPJ supervision of personal-data processing and provision of judicial AI training." ], "relatedIncidents": [], "sources": [ { "label": "BOE — Instruccion 2/2026", "url": "https://www.boe.es/diario_boe/txt.php?id=BOE-A-2026-2205" }, { "label": "CGPJ compendium entry", "url": "https://www.poderjudicial.es/cgpj/es/Temas/Compendio-de-Derecho-Judicial/Instrucciones/Instruccion-2-2026--de-28-de-enero--del-Pleno-del-Consejo-General-del-Poder-Judicial--sobre-la-utilizacion-de-sistemas-de-inteligencia-artificial-en-el-ejercicio-de-la-actividad-jurisdiccional" } ], "country": "ES", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor", "training-cle", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eu-cjeu-ai-ethics-charter", "url": "https://safelegalai.com/regulation/eu/eu-cjeu-ai-ethics-charter", "title": "Artificial Intelligence Ethics Charter", "body": "Court of Justice of the European Union", "bodyShort": "CJEU", "jurisdiction": "eu", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "disclosure", "confidentiality", "supervision", "record-keeping" ], "versions": [ { "version": "Charter adopted by the CJEU", "date": "2026-01-26T00:00:00.000Z", "url": "https://curia.europa.eu/site/upload/docs/application/pdf/2026-03/ethics_charter_ai_-_en.pdf", "changes": "Builds on the CJEU Artificial Intelligence Strategy adopted 19 June 2023 and the work of the AI Management Board." }, { "version": "Artificial Intelligence Strategy", "date": "2023-06-19T00:00:00.000Z", "url": "https://curia.europa.eu/jcms/upload/docs/application/pdf/2023-11/cjeu_ai_strategy.pdf", "changes": "Set the principles and created the AI Management Board composed of Members of both courts and administration representatives." } ], "summary": "A charter adopted by the Court of Justice of the European Union on 26 January 2026 governing AI tools developed or deployed inside the institution, including tools that assist the Court of Justice and the General Court in their judicial functions. It names automation bias as a specific risk to case-law development and judicial quality.", "keyProvisions": [ "'AI systems shall not be entrusted with decision-making responsibilities'; Judges and Advocates General remain in full control of judicial procedures.", "Content generated by AI without verification must be labelled as such if shared beyond the user who requested it.", "Users must take a critical approach to machine-generated output, check accuracy and completeness in each case, and report errors.", "All AI use must comply with the institution's confidentiality protocols, internal security rules and data-protection standards.", "Training, validation and testing datasets must be relevant, representative and, so far as possible, free of error and bias." ], "relatedIncidents": [], "sources": [ { "label": "CJEU — Artificial Intelligence Ethics Charter (PDF)", "url": "https://curia.europa.eu/site/upload/docs/application/pdf/2026-03/ethics_charter_ai_-_en.pdf" }, { "label": "CJEU — Artificial Intelligence", "url": "https://curia.europa.eu/site/jcms/p1_1000081791/en/artificial-intelligence" } ], "country": "EU", "categories": [ "judicial-use", "court-own-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "kr-ai-basic-act-2026", "url": "https://safelegalai.com/regulation/other/kr-ai-basic-act-2026", "title": "Framework Act on the Development of Artificial Intelligence and Establishment of a Foundation for Trust (Law No. 20676)", "body": "National Assembly of the Republic of Korea", "bodyShort": "Korea National Assembly", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification", "disclosure" ], "effectiveDate": "2026-01-22T00:00:00.000Z", "versions": [ { "version": "In force with Enforcement Decree", "date": "2026-01-22T00:00:00.000Z", "url": "https://www.korea.kr/news/policyNewsView.do?newsId=148954629", "changes": "Act and Enforcement Decree in force; fines deferred for at least one year." }, { "version": "Promulgated", "date": "2025-01-21T00:00:00.000Z", "url": "https://www.law.go.kr/lsInfoP.do?lsiSeq=268543" } ], "summary": "Korea's horizontal AI statute, promulgated 21 January 2025 and in force from 22 January 2026 with its Enforcement Decree. It defines high-impact AI by sector, imposes transparency duties including notice that generative outputs are AI-generated, and imposes risk-management, explanation and human-oversight duties on high-impact operators. Justice uses are not separately enumerated.", "keyProvisions": [ "High-impact AI defined by sector with a Ministry of Science and ICT confirmation process.", "Art. 31: advance notice that a service uses AI, and labelling of generative outputs indistinguishable from reality.", "Art. 34: risk-management plan, explainability, user protection, human oversight and documentation for high-impact AI.", "Extraterritorial reach to foreign operators affecting Korean users; penalties deferred at least one year." ], "relatedIncidents": [], "sources": [ { "label": "Act on law.go.kr (official)", "url": "https://www.law.go.kr/lsInfoP.do?lsiSeq=268543" }, { "label": "Government briefing on the Enforcement Decree", "url": "https://www.korea.kr/news/policyNewsView.do?newsId=148954629" }, { "label": "Future of Privacy Forum analysis", "url": "https://fpf.org/blog/south-koreas-new-ai-framework-act-a-balancing-act-between-innovation-and-regulation/" } ], "country": "KR", "categories": [ "ai-act-horizontal", "client-consent-transparency", "evidence-admissibility" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "tw-ai-basic-act-2026", "url": "https://safelegalai.com/regulation/other/tw-ai-basic-act-2026", "title": "Artificial Intelligence Basic Act (Taiwan)", "body": "Legislative Yuan of Taiwan", "bodyShort": "Taiwan Legislative Yuan", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification", "disclosure" ], "effectiveDate": "2026-01-14T00:00:00.000Z", "versions": [ { "version": "Promulgated", "date": "2026-01-14T00:00:00.000Z", "url": "https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=H0160093", "changes": "Promulgated by Presidential Order; 20 articles; NSTC as central competent authority." }, { "version": "Third reading", "date": "2025-12-23T00:00:00.000Z", "url": "https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=H0160093" } ], "summary": "Taiwan's framework AI statute, passed 23 December 2025 and promulgated 14 January 2026. It sets principles for AI development, requires the Ministry of Digital Affairs to establish an internationally aligned risk-classification framework with sectoral tiered rules, requires warnings on high-risk AI products, and obliges government bodies to conduct risk assessments before using AI.", "keyProvisions": [ "Art. 5: high-risk AI products or systems must carry clear notices or warnings.", "Art. 16: MODA to build a risk-classification framework; sector regulators to set tiered management rules.", "Art. 19: government AI use requires risk assessment and mitigation planning.", "National AI strategy committee chaired by the Premier, with the NSTC as central competent authority." ], "relatedIncidents": [], "sources": [ { "label": "Law text (law.moj.gov.tw, official)", "url": "https://law.moj.gov.tw/LawClass/LawAll.aspx?pcode=H0160093" } ], "country": "TW", "categories": [ "ai-act-horizontal", "evidence-admissibility", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "nsw-local-court-generative-ai-practice-note", "url": "https://safelegalai.com/regulation/australia/nsw-local-court-generative-ai-practice-note", "title": "Practice Note: Generative Artificial Intelligence", "body": "Local Court of New South Wales", "bodyShort": "NSW Local Court", "jurisdiction": "australia", "region": "New South Wales", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "prohibition", "disclosure", "verification" ], "effectiveDate": "2026-01-12T00:00:00.000Z", "versions": [ { "version": "Original (commenced 12 January 2026)", "date": "2026-01-12T00:00:00.000Z", "url": "https://localcourt.nsw.gov.au/documents/practice-notes/other/Practice_Note_Generative_Artificial_Intelligence_Commences_12.01.26.pdf" } ], "summary": "A practice note of the Local Court of New South Wales, commencing 12 January 2026, extending the Supreme Court's SC Gen 23 regime to the state's busiest court. Generative AI may not be used to produce the content of affidavits, witness statements or character references, such documents must disclose that AI was not used, and practitioners must verify AI-assisted submissions.", "keyProvisions": [ "Mirrors Supreme Court Practice Note SC Gen 23 for Local Court proceedings.", "Prohibits generative AI in the content of affidavits, witness statements and character references.", "Requires disclosure in evidence documents that generative AI was not used.", "Practitioners remain responsible for verifying citations in AI-assisted submissions." ], "relatedIncidents": [], "article": "nsw-sc-gen-23-strictest-ai-court-rule", "sources": [ { "label": "NSW Local Court — practice note on generative AI (PDF)", "url": "https://localcourt.nsw.gov.au/documents/practice-notes/other/Practice_Note_Generative_Artificial_Intelligence_Commences_12.01.26.pdf" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person", "evidence-admissibility" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "law-society-nsw-solicitors-guide-responsible-ai", "url": "https://safelegalai.com/regulation/australia/law-society-nsw-solicitors-guide-responsible-ai", "title": "A Solicitor's Guide to Responsible Use of Artificial Intelligence", "body": "Law Society of New South Wales", "bodyShort": "Law Society of NSW", "jurisdiction": "australia", "region": "New South Wales", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence", "supervision" ], "effectiveDate": "2026-01-09T00:00:00.000Z", "versions": [ { "version": "January 2026", "date": "2026-01-09T00:00:00.000Z", "url": "https://www.lawsociety.com.au/sites/default/files/2026-01/20260109%20LS4816_PSU_GuidetoAI_2026-01-08a%20FINALv2.pdf", "archiveUrl": "https://web.archive.org/web/20260716080813/https://www.lawsociety.com.au/sites/default/files/2026-01/20260109%20LS4816_PSU_GuidetoAI_2026-01-08a%20FINALv2.pdf" } ], "summary": "A guide published by the Law Society of New South Wales in January 2026 for solicitors on responsible AI use. It explains how the Solicitors Conduct Rules and court practice notes, including SC Gen 23, apply to AI-assisted work, and sets expectations on verification, confidentiality, client communication, supervision and selecting tools with appropriate data protections.", "keyProvisions": [ "Solicitors remain responsible for all AI-assisted work and must verify output.", "Confidential information must not be entered into tools without appropriate protections.", "Explains compliance with NSW court practice notes on generative AI, including affidavit restrictions.", "Firms should adopt AI policies, training and supervision arrangements." ], "relatedIncidents": [], "article": "australia-law-society-ai-guidance-history", "sources": [ { "label": "Law Society of NSW — Solicitor's guide to responsible use of AI (PDF)", "url": "https://www.lawsociety.com.au/sites/default/files/2026-01/20260109%20LS4816_PSU_GuidetoAI_2026-01-08a%20FINALv2.pdf" } ], "country": "AU", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-co-rule-change-2026-02-ai-rpc-amendments", "url": "https://safelegalai.com/regulation/us-state/us-co-rule-change-2026-02-ai-rpc-amendments", "title": "Colorado Rule Change 2026(02): AI amendments to the Colorado Rules of Professional Conduct", "body": "Colorado Supreme Court", "bodyShort": "Colorado Supreme Court", "jurisdiction": "us-state", "region": "Colorado", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "supervision", "confidentiality" ], "effectiveDate": "2026-01-08T00:00:00.000Z", "versions": [ { "version": "Rule Change 2026(02)", "date": "2026-01-08T00:00:00.000Z", "url": "https://www.coloradojudicial.gov/sites/default/files/2026-01/Rule%20Change%202026(02).pdf", "changes": "Amended and adopted en banc, effective immediately" } ], "summary": "Colorado became the first US jurisdiction to amend its lawyer ethics rules specifically for AI. Scope [20A] confirms technology does not diminish a lawyer's responsibilities and that misuse is disciplinable; Comment [9] to Rule 1.1 enumerates the rules AI implicates and preserves the duty of independent judgment.", "keyProvisions": [ "Scope [20A]: technology including artificial intelligence does not diminish a lawyer's responsibilities under the Rules, and a lawyer who uses technology directly or indirectly may be disciplined for a resulting violation", "Comment [8] to Rule 1.1: lawyers must keep abreast of the benefits and risks associated with relevant technology", "Comment [9] to Rule 1.1: AI use can implicate Rules 1.4, 1.5, 1.6, 3.1, 3.3, 5.1, 5.2, 5.3, 7.1 and 8.4(g)", "Comment [9]: reliance on technology does not diminish the lawyer's duty to exercise independent judgment", "Adopted en banc 8 January 2026, effective immediately" ], "relatedIncidents": [], "sources": [ { "label": "Rule Change 2026(02) (PDF)", "url": "https://www.coloradojudicial.gov/sites/default/files/2026-01/Rule%20Change%202026(02).pdf" }, { "label": "Colorado Lawyer analysis (CBA official publication)", "url": "https://cl.cobar.org/features/the-2026-ai-related-amendments-to-the-colorado-rules-of-professional-conduct/" } ], "country": "US", "categories": [ "bar-guidance", "competence-duty", "confidentiality-client-data", "fees-billing", "supervision", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "qicdrc-practice-direction-1-2026-ai", "url": "https://safelegalai.com/regulation/gulf/qicdrc-practice-direction-1-2026-ai", "title": "Practice Direction No. 1 of 2026 and Practice Guidance on the Use of Artificial Intelligence", "body": "Qatar International Court and Dispute Resolution Centre", "bodyShort": "QICDRC", "jurisdiction": "gulf", "region": "Qatar", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure" ], "effectiveDate": "2026-01-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2026-01-06T00:00:00.000Z", "url": "https://www.qicdrc.gov.qa/media-center/news/qicdrc-issues-practice-direction-no-1-2026-and-practice-guidance-use-artificial", "archiveUrl": "https://web.archive.org/web/20260904134039/https://www.qicdrc.gov.qa/media-center/news/qicdrc-issues-practice-direction-no-1-2026-and-practice-guidance-use-artificial" } ], "summary": "A practice direction and accompanying practice guidance issued by the Qatar International Court and Dispute Resolution Centre on 6 January 2026 on the use of AI in proceedings before the Qatar Financial Centre courts. Parties and representatives must verify AI-assisted content, may be required to disclose AI use, and remain responsible for fabricated or inaccurate material.", "keyProvisions": [ "Parties and legal representatives are responsible for the accuracy of AI-assisted submissions.", "The court may require disclosure of how AI was used in preparing documents.", "Follows a 2025 QFC court judgment addressing fabricated AI-generated citations." ], "relatedIncidents": [ "sheppard-v-jillion-qfc" ], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "QICDRC — news release on Practice Direction No. 1 of 2026", "url": "https://www.qicdrc.gov.qa/media-center/news/qicdrc-issues-practice-direction-no-1-2026-and-practice-guidance-use-artificial" } ], "country": "QA", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "au-nsw-pic-procedural-direction-13-generative-ai", "url": "https://safelegalai.com/regulation/australia/au-nsw-pic-procedural-direction-13-generative-ai", "title": "Procedural Direction PIC13: Use of Generative Artificial Intelligence", "body": "Personal Injury Commission of New South Wales", "bodyShort": "NSW Personal Injury Commission", "jurisdiction": "australia", "region": "New South Wales", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties", "providers" ], "requirements": [ "prohibition", "disclosure", "verification", "confidentiality", "record-keeping" ], "effectiveDate": "2026-01-01T00:00:00.000Z", "versions": [ { "version": "Commenced 1 January 2026", "date": "2026-01-01T00:00:00.000Z", "url": "https://www.pi.nsw.gov.au/resources/procedural-directions/procedural-direction-pic13-use-of-generative-artificial-intelligence" } ], "summary": "The Personal Injury Commission's direction, in force from 1 January 2026, applies the New South Wales model to workers compensation and motor accidents disputes and goes further in two respects: expert witnesses must disclose AI use, program and version and keep prompt records, and parties may not use AI technology during a hearing.", "keyProvisions": [ "Gen AI must not generate the content of written statements; statements must disclose that Gen AI was not used to create their content.", "Confidential, de-identified or legally privileged material must not be entered into a Gen AI program unless the platform is contractually closed and does not train on the data.", "Where Gen AI assisted written submissions, the author must personally verify that all citations, authorities and legislative references exist and are accurate; verification may not be done with an AI tool.", "Experts must disclose any Gen AI use, the program and version, how it was applied, and keep records of prompts or settings.", "Parties must not use AI technology, including speech-to-text and transcription applications, during Commission hearings whether in person or virtual." ], "relatedIncidents": [], "sources": [ { "label": "Personal Injury Commission — Procedural Direction PIC13", "url": "https://www.pi.nsw.gov.au/resources/procedural-directions/procedural-direction-pic13-use-of-generative-artificial-intelligence" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "evidence-admissibility", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-sk-kb-ga-pd-12-artificial-intelligence", "url": "https://safelegalai.com/regulation/canada/ca-sk-kb-ga-pd-12-artificial-intelligence", "title": "General Application Practice Directive #12: Use of Artificial Intelligence in Court Submissions", "body": "Court of King's Bench for Saskatchewan", "bodyShort": "Saskatchewan King's Bench", "jurisdiction": "canada", "region": "Saskatchewan", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "supervision", "competence", "permissive" ], "versions": [ { "version": "GA-PD #12", "date": "2026-01-01T00:00:00.000Z", "url": "https://sasklawcourts.ca/wp-content/uploads/2026/05/KB_GA-PD-12.pdf" } ], "summary": "Signed by Chief Justice M.D. Popescul and effective 1 January 2026. The directive does not prohibit appropriate AI use; it places the onus on whoever files or presents materials to ensure accuracy, authenticity and reliability, and requires independent verification of every AI-derived authority against official, commercial or CanLII sources.", "keyProvisions": [ "Nothing in the directive prohibits appropriate use of AI; the onus is on the filer for accuracy, authenticity and reliability.", "Counsel and self-represented litigants are responsible for all information, authorities and analysis in written and oral submissions; reliance on AI does not diminish that duty.", "AI-generated, summarised or cited legal material must be independently verified against official court or legislative sites, recognised commercial publishers, or CanLII.", "Verification must confirm that citations correspond to real authorities, quotations are accurate and AI summaries reflect the source.", "Unverified or fabricated authorities may attract cost consequences or other sanctions; blind reliance on AI is inconsistent with professional competence." ], "relatedIncidents": [], "sources": [ { "label": "Saskatchewan Courts — GA-PD #12 (official PDF)", "url": "https://sasklawcourts.ca/wp-content/uploads/2026/05/KB_GA-PD-12.pdf" }, { "label": "Saskatchewan Courts — announcement of two new practice directives", "url": "https://sasklawcourts.ca/court-of-kings-bench-issues-two-new-practice-directives/" } ], "country": "CA", "categories": [ "verification-duty", "litigant-in-person", "competence-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-az-judicial-conduct-rule-2-5-technology-competence", "url": "https://safelegalai.com/regulation/us-state/us-az-judicial-conduct-rule-2-5-technology-competence", "title": "Arizona Code of Judicial Conduct Rule 2.5, Comment 1: judicial technology competence", "body": "Supreme Court of Arizona", "bodyShort": "Arizona Supreme Court", "jurisdiction": "us-state", "region": "Arizona", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "competence" ], "effectiveDate": "2026-01-01T00:00:00.000Z", "versions": [ { "version": "Comment 1 as amended", "date": "2026-01-01T00:00:00.000Z", "url": "https://azcjc.azcourts.gov/Portals/5/137/rules/2025%20Code%20of%20Judicial%20Conduct.pdf", "changes": "Adopted 27 August 2025, effective 1 January 2026" } ], "summary": "Arizona added an explicit technology-competence duty to its Code of Judicial Conduct. Comment 1 to Rule 2.5 now requires judges to have the legal knowledge, skill, thoroughness and preparation reasonably necessary to perform judicial responsibilities, including use and knowledge of the benefits and risks of relevant technology.", "keyProvisions": [ "Judicial competence under Rule 2.5 expressly includes the use of, and knowledge of the benefits and risks associated with, relevant technology", "Effective 1 January 2026", "Parallels the attorney technology-competence duty under Rule 42, ER 1.1, Comment 6" ], "relatedIncidents": [], "sources": [ { "label": "Arizona Code of Judicial Conduct 2025 (PDF)", "url": "https://azcjc.azcourts.gov/Portals/5/137/rules/2025%20Code%20of%20Judicial%20Conduct.pdf" } ], "country": "US", "categories": [ "judicial-use", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-fl-17th-circuit-ao-2026-03-gen", "url": "https://safelegalai.com/regulation/us-state/us-fl-17th-circuit-ao-2026-03-gen", "title": "Florida 17th Judicial Circuit Administrative Order 2026-03-GEN: Use of AI in Court Filings", "body": "Seventeenth Judicial Circuit of Florida (Broward County)", "bodyShort": "FL 17th Judicial Circuit", "jurisdiction": "us-state", "region": "Florida", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "verification" ], "versions": [ { "version": "AO 2026-03-GEN", "date": "2026-01-01T00:00:00.000Z", "url": "https://www.17th.flcourts.org/wp-content/uploads/2026/01/AO2026-03-GEN-Use-of-AI-in-Court-Filings.pdf", "changes": "Initial administrative order; the source gives a month only (issued January 2026), so the day is recorded as the 1st" } ], "summary": "The Seventeenth Judicial Circuit of Florida issued an administrative order in January 2026 on the use of artificial intelligence in court filings. Several other Florida circuits have parallel orders. The operative requirements were not read in this run and must be confirmed.", "keyProvisions": [ "Circuit-level administrative order governing AI use in court filings (terms not yet verified)", "Parallel orders exist in the 11th, 15th and 19th Judicial Circuits" ], "relatedIncidents": [], "sources": [ { "label": "AO 2026-03-GEN (PDF)", "url": "https://www.17th.flcourts.org/wp-content/uploads/2026/01/AO2026-03-GEN-Use-of-AI-in-Court-Filings.pdf" }, { "label": "11th Judicial Circuit AI administrative order", "url": "https://www.jud11.flcourts.org/docs/ao-html/1b10ef43-b414-40e5-88d7-401f46d0a081.html" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-mt-13th-district-local-rule-35-ai", "url": "https://safelegalai.com/regulation/us-state/us-mt-13th-district-local-rule-35-ai", "title": "Montana Thirteenth Judicial District Court Local Rule 35 — Use of Generative Artificial Intelligence", "body": "Montana Thirteenth Judicial District Court (Yellowstone County)", "bodyShort": "MT 13th Judicial District Court", "jurisdiction": "us-state", "region": "Montana", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "certification", "verification" ], "effectiveDate": "2026-01-01T00:00:00.000Z", "versions": [ { "version": "Local rules signed 1 January 2026", "date": "2026-01-01T00:00:00.000Z", "url": "https://www.yellowstonecountymt.gov/clerk_court/LocalRules.pdf", "changes": "Local Rule 35 on generative AI; some practitioner summaries describe an earlier version effective 1 April 2025" } ], "summary": "Local Rule 35 of Montana's Thirteenth Judicial District Court (Yellowstone County) requires an attorney or self-represented party who used generative AI in preparing any document filed with the court to identify the type of AI used and to certify that every citation to law or to the record has been verified as accurate.", "keyProvisions": [ "Applies to attorneys and self-represented parties filing in the 13th Judicial District Court", "The filer must identify the type of generative AI used", "The filer must certify that every citation to the law or the record has been verified as accurate" ], "relatedIncidents": [], "sources": [ { "label": "Yellowstone County District Court local rules (official PDF, unreachable from this environment)", "url": "https://www.yellowstonecountymt.gov/clerk_court/LocalRules.pdf" }, { "label": "Montana LawHelp, AI and Legal Resources", "url": "https://www.montanalawhelp.org/resource/ai-and-legal-resources" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "sa-courts-generative-ai-litigation-guidelines", "url": "https://safelegalai.com/regulation/australia/sa-courts-generative-ai-litigation-guidelines", "title": "Guidelines concerning the use of Generative Artificial Intelligence in litigation in South Australian Courts", "body": "Courts Administration Authority of South Australia", "bodyShort": "SA Courts", "jurisdiction": "australia", "region": "South Australia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure", "confidentiality" ], "effectiveDate": "2026-01-01T00:00:00.000Z", "versions": [ { "version": "Published 24 December 2025, effective 1 January 2026", "date": "2025-12-24T00:00:00.000Z", "url": "https://www.courts.sa.gov.au/2025/12/24/guidelines-concerning-the-use-of-generative-artificial-intelligence-in-litigation-in-south-australian-courts/" } ], "summary": "Guidelines published by the Courts Administration Authority of South Australia on 24 December 2025 and effective 1 January 2026, applying to all South Australian courts operating under the Uniform Civil, Special Statutory and Joint Criminal Rules. Parties must verify AI-assisted material, protect confidential information, and the rules are to be amended to give the guidelines effect.", "keyProvisions": [ "Applies across all South Australian courts under the UCR, USSR and JCR.", "Practitioners and parties must verify any AI-assisted content before filing.", "Confidential and suppressed information must not be entered into public AI tools.", "Accompanied by an information sheet for self-represented litigants; court rules to be amended to incorporate the guidelines." ], "relatedIncidents": [], "sources": [ { "label": "Courts Administration Authority of South Australia — guidelines announcement", "url": "https://www.courts.sa.gov.au/2025/12/24/guidelines-concerning-the-use-of-generative-artificial-intelligence-in-litigation-in-south-australian-courts/" } ], "country": "AU", "categories": [ "verification-duty", "litigant-in-person", "confidentiality-client-data", "disclosure-filings" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-nyc-bar-formal-opinion-2025-6", "url": "https://safelegalai.com/regulation/us-state/us-nyc-bar-formal-opinion-2025-6", "title": "New York City Bar Formal Opinion 2025-6: Ethical Issues Affecting Use of AI to Record, Transcribe, and Summarize Conversations with Clients", "body": "New York City Bar Association Committee on Professional Ethics", "bodyShort": "NYC Bar", "jurisdiction": "us-state", "region": "New York", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "consent", "confidentiality", "verification", "competence", "supervision" ], "versions": [ { "version": "Original", "date": "2025-12-22T00:00:00.000Z", "url": "https://www.nycbar.org/reports/formal-opinion-2025-6/", "changes": "Initial publication" } ], "summary": "Formal Opinion 2025-6 addresses AI note-takers and transcription tools used in client communications. Clients must be notified and consent obtained whenever an AI-empowered system records their calls; lawyers must independently review AI transcripts and summaries, understand the tools' technical limits under Rule 1.1, and manage privilege and discovery risks in stored records.", "keyProvisions": [ "Clients must be notified and their consent obtained whenever calls are recorded by an AI-empowered system", "Lawyers must independently verify the accuracy of AI-generated transcripts and summaries rather than relying on them unreviewed", "Attorneys must safeguard AI-generated records and consider privilege, storage and discovery implications", "Rule 1.1 competence requires understanding the technical features and limits of the AI tools used by the lawyer and by clients", "Where clients use their own AI tools, retainers should disclaim reliance on unreviewed summaries and warn of confidentiality risks", "Firms must train subordinates on ethical duties for AI in client communications" ], "relatedIncidents": [], "sources": [ { "label": "NYC Bar Formal Opinion 2025-6", "url": "https://www.nycbar.org/reports/formal-opinion-2025-6/" } ], "country": "US", "categories": [ "bar-guidance", "confidentiality-client-data", "client-consent-transparency", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cepej-guidelines-generative-ai-courts", "url": "https://safelegalai.com/regulation/international/cepej-guidelines-generative-ai-courts", "title": "Guidelines on the use of generative AI by courts and tribunals (CEPEJ(2025)18Final)", "body": "European Commission for the Efficiency of Justice, Council of Europe", "bodyShort": "CEPEJ", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "disclosure", "verification", "confidentiality" ], "effectiveDate": "2025-12-19T00:00:00.000Z", "versions": [ { "version": "Final", "date": "2025-12-19T00:00:00.000Z", "url": "https://rm.coe.int/cepej-2025-18final-en-draft-guidelines-on-the-use-of-generative-ai-for/48802a4ad1" } ], "summary": "Guidelines adopted 19 December 2025 by the Council of Europe's CEPEJ for courts and tribunals in its 46 member states on using generative AI. Judicial use must be transparent, AI output can never bind a judge, the right to a human judge is guaranteed, and confidential case data must be protected. They replace the February 2024 information note.", "keyProvisions": [ "A human judge must always take the decision; generative AI output is never binding.", "Judicial use of generative AI should be transparent to parties where it affects proceedings.", "Confidential and personal data in case files must not be entered into unsecured tools.", "Outputs must be verified for accuracy before any reliance." ], "relatedIncidents": [], "sources": [ { "label": "Council of Europe — CEPEJ(2025)18Final (PDF)", "url": "https://rm.coe.int/cepej-2025-18final-en-draft-guidelines-on-the-use-of-generative-ai-for/48802a4ad1" } ], "country": "INT", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "verification-duty", "confidentiality-client-data", "disclosure-filings", "data-protection" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-hi-supreme-court-ai-committee-final-report", "url": "https://safelegalai.com/regulation/us-state/us-hi-supreme-court-ai-committee-final-report", "title": "Hawaii Supreme Court Committee on Artificial Intelligence and the Courts: Final Report (SCMF-24-0000305)", "body": "Supreme Court of Hawaii, Committee on Artificial Intelligence and the Courts", "bodyShort": "Hawaii Supreme Court", "jurisdiction": "us-state", "region": "Hawaii", "type": "report", "status": "proposed", "appliesTo": [ "judges", "court-staff", "lawyers", "litigants" ], "requirements": [ "disclosure", "verification" ], "versions": [ { "version": "Final Report", "date": "2025-12-16T00:00:00.000Z", "url": "https://www.courts.state.hi.us/", "changes": "Filed in SCMF-24-0000305" } ], "summary": "Hawaii's judiciary studied AI for eighteen months and concluded HRCP Rule 11 already covers AI abuse, while recommending the Supreme Court consider a federal-style AI disclosure declaration for state-court filings. The committee separately adopted internal guardrails for judiciary use of AI tools.", "keyProvisions": [ "Existing rules including HRCP Rule 11 are broad enough to provide adequate safeguards against AI abuse in state-court filings", "The Court should consider adopting a disclosure rule mirroring D. Haw. General Order 23-1", "Under the proposed rule an AI-assisted filing would be accompanied by a declaration confirming AI use and that all cited authorities have been verified", "The committee adopted Hawaii State Judiciary Guardrails for the Acceptable Use of AI Tools and Platforms for judges and staff", "Advisory only; no implementing Supreme Court order as of September 2026" ], "relatedIncidents": [], "sources": [ { "label": "Hawaii State Judiciary (official site)", "url": "https://www.courts.state.hi.us/" } ], "country": "US", "categories": [ "disclosure-filings", "judicial-use", "court-own-use", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "fr-conseil-etat-charte-ia-juridiction-administrative", "url": "https://safelegalai.com/regulation/other/fr-conseil-etat-charte-ia-juridiction-administrative", "title": "Charte d'utilisation de l'intelligence artificielle au sein de la juridiction administrative", "body": "Conseil d'État (Council of State), France", "bodyShort": "Conseil d'État", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "competence", "prohibition", "supervision" ], "versions": [ { "version": "First version", "date": "2025-12-11T00:00:00.000Z", "url": "https://www.conseil-etat.fr/Media/actualites/documents/2025/decembre-2025/charte-d-utilisation-de-l-intelligence-artificielle-au-sein-de-la-juridiction-administrative", "changes": "Initial charter, adopted after consultation with staff, unions and elected representatives and reviewed by the collège de déontologie." } ], "summary": "A deontological charter published on 11 December 2025 binding every member, magistrate and staff member of the French administrative jurisdiction — Conseil d'État, cours administratives d'appel, tribunaux administratifs, Cour nationale du droit d'asile and Tribunal du stationnement payant. It sets seven principles for generative AI, excludes high-risk AI systems, and covers external chatbots, which are not blocked but tightly conditioned.", "keyProvisions": [ "'L'IA est un outil qui jamais ne décide': AI may not interpret a rule of law or case law, establish or assess facts, apply law to a case, reason legally or propose an outcome.", "Human verification is systematic — AI output, especially legal and case-law references, must never be assumed true or complete and must be checkable by another means.", "Users adopt AI-generated content as their own and cannot 'se défausser sur l'IA' for errors or omissions.", "Give up using external AI where data security and confidentiality cannot be assured: entering information into a public chatbot 'équivaut à la publier sur internet'.", "Users must guard against anchoring, confirmation, training and discriminatory bias, and weigh the environmental cost of each prompt.", "The charter covers only limited-risk AI under the AI Act; high-risk judicial AI (Annex III point 8(a)) is excluded." ], "relatedIncidents": [], "sources": [ { "label": "Conseil d'État — charter page (official)", "url": "https://www.conseil-etat.fr/qui-sommes-nous/deontologie/charte-d-utilisation-de-l-intelligence-artificielle-au-sein-de-la-juridiction-administrative" }, { "label": "Conseil d'État — charter (official PDF)", "url": "https://www.conseil-etat.fr/Media/actualites/documents/2025/decembre-2025/charte-d-utilisation-de-l-intelligence-artificielle-au-sein-de-la-juridiction-administrative" } ], "country": "FR", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "court-own-use", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-unesco-guidelines-ai-courts-tribunals", "url": "https://safelegalai.com/regulation/international/int-unesco-guidelines-ai-courts-tribunals", "title": "UNESCO Guidelines for the use of AI systems in courts and tribunals", "body": "United Nations Educational, Scientific and Cultural Organization", "bodyShort": "UNESCO", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence", "risk-classification" ], "effectiveDate": "2025-12-03T00:00:00.000Z", "versions": [ { "version": "Final", "date": "2025-12-03T00:00:00.000Z", "url": "https://unesdoc.unesco.org/ark:/48223/pf0000396582", "changes": "Published 3 December 2025; launched at the Athens Roundtable on AI and the Rule of Law, London, 4 December 2025." }, { "version": "Draft for consultation", "date": "2024-09-01T00:00:00.000Z", "url": "https://unesdoc.unesco.org/ark:/48223/pf0000390781" } ], "summary": "Published 3 December 2025 after a global consultation, the Guidelines are UNESCO's ethical and operational framework for AI in judicial systems. Fifteen principles in five groups - human rights and equality; safety, security and accuracy; explainability, transparency and auditability; awareness and accountability; human-centred decision-making - are paired with recommendations for judicial organisations and individual judges.", "keyProvisions": [ "AI is an assistive, not substitutive, tool; it shall not replace judicial decision-making (principles 1.13-1.15).", "Judges should verify AI-generated information independently, avoid over-reliance and disclose when AI tools influenced a decision.", "Judicial organisations should set procurement standards with risk and rights-impact review, run training and monitor systems across their lifecycle, including the ability to disable them.", "Explainability: a decision that cannot be understood cannot be used to affect someone's rights." ], "relatedIncidents": [], "sources": [ { "label": "UNESCO — Guidelines announcement", "url": "https://www.unesco.org/en/articles/guidelines-use-ai-systems-courts-and-tribunals" }, { "label": "UNESCO — AI in the courtroom (launch)", "url": "https://www.unesco.org/en/articles/ai-courtroom-unescos-new-guidelines-judiciary" }, { "label": "UNESDOC record", "url": "https://unesdoc.unesco.org/ark:/48223/pf0000396582" } ], "country": "INT", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor", "training-cle", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cn-beijing-tongzhou-court-ai-fabricated-cases-2025", "url": "https://safelegalai.com/regulation/other/cn-beijing-tongzhou-court-ai-fabricated-cases-2025", "title": "Beijing Tongzhou District People's Court: counsel criticised for filing AI-fabricated precedents", "body": "Beijing Tongzhou District People's Court, selected for the People's Court Case Database", "bodyShort": "Tongzhou District Court", "jurisdiction": "other", "region": "Beijing", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties", "litigants" ], "requirements": [ "verification" ], "effectiveDate": "2025-12-01T00:00:00.000Z", "versions": [ { "version": "Case Database entry", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.chinanews.com.cn/fz/2026/01-15/10552373.shtml", "changes": "Selected for the People's Court Case Database in December 2025; reported nationally 15-16 January 2026." } ], "summary": "A plaintiff's lawyer filed supplementary submissions citing two reference cases generated by an AI model and never verified. The judge found the cited case numbers belonged to unrelated real cases and wrote a criticism into the judgment without a fine. Entered in the People's Court Case Database in December 2025, it sets graduated consequences for unverified AI citations.", "keyProvisions": [ "Lawyers must verify the authenticity of AI-sourced materials before submission.", "Consequences graduated by severity: criticism and education for minor cases; fines or detention under Civil Procedure Law art. 114 for serious cases.", "Applies to all litigation participants, including self-represented parties." ], "relatedIncidents": [], "sources": [ { "label": "China News Service report", "url": "https://www.chinanews.com.cn/fz/2026/01-15/10552373.shtml" }, { "label": "Ministry of Justice legal-publicity repost", "url": "http://legalinfo.moj.gov.cn/pub/sfbzhfx/zhfxfzzx/fzzxyw/202601/t20260116_530661.html" }, { "label": "Beijing Daily report", "url": "https://xinwen.bjd.com.cn/content/s6968e6add5de6b42d6c2aea2.html" } ], "country": "CN", "categories": [ "verification-duty", "sanctions-record", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cz-usoud-i-us-3004-25-fabricated-citations", "url": "https://safelegalai.com/regulation/other/cz-usoud-i-us-3004-25-fabricated-citations", "title": "Ústavní soud, usnesení sp. zn. I. ÚS 3004/25 — procedural fine for a constitutional complaint citing non-existent decisions", "body": "Constitutional Court of the Czech Republic [Ústavní soud]", "bodyShort": "Ústavní soud", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "verification" ], "versions": [ { "version": "1.0", "date": "2025-12-01T00:00:00.000Z", "url": "https://nalus.usoud.cz/Search/GetText.aspx?sz=1-3004-25_1", "changes": "Decision imposing a CZK 25,000 procedural fine on counsel." } ], "summary": "The Czech Constitutional Court fined counsel Mgr. Pavel Kehl CZK 25,000 over a constitutional complaint citing decisions that did not exist or whose holdings were grossly misdescribed. The court identified anglicisms, sentence fragments and inconsistent citation formatting as indicators of AI drafting, and held counsel fully responsible for hallucinated arguments.", "keyProvisions": [ "Cited authorities including III. ÚS 2987/18, III. ÚS 3881/18 and II. ÚS 2264/14 either did not exist or were grossly misinterpreted.", "Indicators of AI generation: anglicisms, incomplete sentence fragments, inconsistent citation formatting and confusion between decision types.", "A lawyer bears full responsibility for every argument in a filing, including any hallucinating arguments produced by artificial intelligence.", "The CZK 25,000 fine was a quarter of the CZK 100,000 statutory maximum, reflecting a first offence and the remediable nature of the defects." ], "relatedIncidents": [], "sources": [ { "label": "Ústavní soud — I. ÚS 3004/25 full text (NALUS)", "url": "https://nalus.usoud.cz/Search/GetText.aspx?sz=1-3004-25_1" } ], "country": "CZ", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "fi-bar-tietoturvaopas-b052", "url": "https://safelegalai.com/regulation/other/fi-bar-tietoturvaopas-b052", "title": "B 05.2 Data Security Manual [Tietoturvaopas], Annex 5 'AI guide for the advocate' [Tekoälyopas asianajajalle]", "body": "Finnish Bar Association [Suomen Asianajajaliitto / Suomen Asianajajat], board", "bodyShort": "Finnish Bar Association", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "record-keeping", "risk-classification" ], "versions": [ { "version": "13.11.2025 update", "date": "2025-12-01T00:00:00.000Z", "url": "https://asianajajat.fi/wp-content/uploads/2025/12/B-05.2-Tietoturvaopas-hallitus-23.11.2018-paiv.-12.12.2019-24.9.2021-14.12.2023-19.4.2024-ja-13.11.2025-voimassa-1.12.2025-lukien.pdf", "changes": "Board updated Annex 5, the AI guide, on 13 November 2025; the manual is in force from 1 December 2025. Annex 5 now runs to twelve numbered requirements plus a Microsoft 365 Copilot deployment guide." }, { "version": "19.4.2024 update", "date": "2024-04-19T00:00:00.000Z", "url": "https://asianajajat.fi/asianajajalle/saantely/saadoksia-ja-ohjeita/", "changes": "Board added Annex 5 (the AI guide) and amended the 'Background' section to note that AI use increases the risks of handling confidential data." }, { "version": "Original", "date": "2018-11-23T00:00:00.000Z", "url": "https://asianajajat.fi/asianajajalle/saantely/saadoksia-ja-ohjeita/", "changes": "Manual adopted by the Bar's board on 23 November 2018; earlier updates 12 Dec 2019, 24 Sept 2021 and 14 Dec 2023." } ], "summary": "The Finnish Bar's information-security manual, recommendatory in character but functioning as commentary on the binding B 05.1 guideline. Its Annex 5, first added in April 2024 and substantially updated in November 2025, is Finland's principal profession-facing generative-AI guidance for advocates, in force from 1 December 2025.", "keyProvisions": [ "Annex 5 point 2: whether advice rests wholly or partly on AI output, the advocate is personally liable to the client; tools err in ways that are hard to detect and outputs can look credible while containing factual errors or fabricated references, so the advocate must always verify correctness before use.", "Annex 5 point 1: a business-grade licence of the AI application must be acquired, applying binding guideline B 05.1 point 6; a free consumer version is not a business subscription.", "Annex 5 point 4: AI tools may be used only if it can be assured that privileged material is not used, stored or shared contrary to confidentiality; terms and settings, including those of extensions and agents, must be checked so inputs are not used for model training or to answer other users.", "Annex 5 point 6: firms must ensure staff AI literacy under AI Act Art. 4, covering how AI works, its limits, prohibited uses, human oversight (outputs assessed critically before use) and transparency about when AI was used and how it shaped the result; literacy must be refreshed as systems change.", "Annex 5 points 3, 7, 8 and 9: periodic risk assessment of each tool; access rights scoped so AI cannot surface material beyond its intended audience; blocking use of firm data or user content for third-party model training; and logging and monitoring of AI use assigned to competent staff because models drift and fabricate.", "Main text: where a cloud service such as an AI solution offers a version with wider security features that separate the customer's data from other users' data, that version must be adopted." ], "relatedIncidents": [], "sources": [ { "label": "Suomen Asianajajat — B 05.2 Tietoturvaopas (PDF, in force 1 Dec 2025)", "url": "https://asianajajat.fi/wp-content/uploads/2025/12/B-05.2-Tietoturvaopas-hallitus-23.11.2018-paiv.-12.12.2019-24.9.2021-14.12.2023-19.4.2024-ja-13.11.2025-voimassa-1.12.2025-lukien.pdf" }, { "label": "Suomen Asianajajat — regulations and guidelines index", "url": "https://asianajajat.fi/asianajajalle/saantely/saadoksia-ja-ohjeita/" } ], "country": "FI", "categories": [ "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "bar-guidance", "training-cle", "procurement-vendor", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hk-digital-policy-office-generative-ai-guideline", "url": "https://safelegalai.com/regulation/hong-kong/hk-digital-policy-office-generative-ai-guideline", "title": "Hong Kong Generative AI Technical and Application Guideline", "body": "Digital Policy Office, Government of Hong Kong", "bodyShort": "HK Digital Policy Office", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification", "permissive" ], "effectiveDate": "2025-04-01T00:00:00.000Z", "versions": [ { "version": "December 2025 update", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf", "changes": "Updated edition of the April 2025 guideline." }, { "version": "April 2025", "date": "2025-04-01T00:00:00.000Z", "url": "https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf" } ], "summary": "A cross-sector governance guideline published by Hong Kong's Digital Policy Office in April 2025 and updated in December 2025. It sets out a technical and application framework for developers and deployers of generative AI, covering governance, risk assessment, data protection and safety. It is not legal-sector specific but applies to law firms deploying AI in Hong Kong.", "keyProvisions": [ "Provides a governance and risk framework for generative AI developers and users.", "Addresses data protection, security and content accuracy across sectors.", "Voluntary guideline forming part of the Ethical AI Framework." ], "relatedIncidents": [], "sources": [ { "label": "Digital Policy Office — Generative AI guideline (PDF)", "url": "https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf" } ], "country": "HK", "categories": [ "ai-act-horizontal", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "hk-law-society-circular-25-824-generative-ai", "url": "https://safelegalai.com/regulation/hong-kong/hk-law-society-circular-25-824-generative-ai", "title": "Circular 25-824: Generative AI in Legal Practice: Risks and Tips", "body": "Law Society of Hong Kong", "bodyShort": "Law Society of Hong Kong", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence" ], "effectiveDate": "2025-12-01T00:00:00.000Z", "versions": [ { "version": "December 2025", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.hklawsoc.org.hk/-/media/HKLS/pub_e/circular/2025/25-824a1.pdf" } ], "summary": "A circular issued by the Law Society of Hong Kong in December 2025 setting out the risks of generative AI in legal practice and practical tips for solicitors. It cautions that output may be inaccurate or fabricated and must be verified, that client confidentiality must be protected when using tools, and that solicitors remain responsible for all work product.", "keyProvisions": [ "Solicitors must verify AI-generated content, including citations, before use.", "Client confidential information must not be entered into tools without safeguards.", "Solicitors should understand the tool and its data handling before adoption." ], "relatedIncidents": [], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "Law Society of Hong Kong — Circular 25-824 (PDF)", "url": "https://www.hklawsoc.org.hk/-/media/HKLS/pub_e/circular/2025/25-824a1.pdf" } ], "country": "HK", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "nl-nova-aanbevelingen-ai-advocatuur", "url": "https://safelegalai.com/regulation/other/nl-nova-aanbevelingen-ai-advocatuur", "title": "Aanbevelingen AI in de advocatuur", "body": "Nederlandse orde van advocaten (Netherlands Bar)", "bodyShort": "NOvA", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "consent", "competence", "supervision", "permissive" ], "versions": [ { "version": "First version", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.advocatenorde.nl/voor-advocaten/digitalisering-ai/aanbevelingen-ai-de-advocatuur", "changes": "First NOvA AI recommendations, prepared by the projectgroep Digitalisering & AI; reported by Het Advocatenblad on 10 December 2025. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "The Dutch Bar's recommendations on responsible AI use, published at the end of 2025 and organised around the five statutory core values of the profession. They are notable in Europe for recommending that lawyers ask the client's permission before using AI on a file, and for requiring manual verification of every citation and fact.", "keyProvisions": [ "Deskundigheid: basic knowledge of generative AI is essential for every lawyer — LLM principles, prompting, bias mitigation, hallucinations, error detection, regulation and cybersecurity.", "'Verifieer altijd de output': check citations, case law and facts manually before use, and use only tools that cite sources.", "'Vraag toestemming aan de cliënt voor AI-gebruik in het dossier' — seek client permission for AI use on the matter.", "No confidential or client data in free or public AI models; apply privacy by design, run a DPIA, and keep input and output inside the firm environment.", "Know the data flows — storage country, sub-processors, security — and select suppliers by reading the actual contract and privacy terms, not marketing claims.", "Onafhankelijkheid: supervise staff and suppliers, adopt a firm-wide AI policy and implement a control cycle; be alert to tools that simply confirm the prompt." ], "relatedIncidents": [], "sources": [ { "label": "NOvA — Aanbevelingen AI in de advocatuur (official)", "url": "https://www.advocatenorde.nl/voor-advocaten/digitalisering-ai/aanbevelingen-ai-de-advocatuur" }, { "label": "Het Advocatenblad — AI-richtlijn: advocaat blijft eindverantwoordelijk, 10 Dec 2025 (secondary, for the date)", "url": "https://www.advocatenblad.nl/2025/12/10/ai-richtlijn-advocaat-blijft-eindverantwoordelijk/" } ], "country": "NL", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ny-ucs-ai-advisory-committee-2025-report", "url": "https://safelegalai.com/regulation/us-state/ny-ucs-ai-advisory-committee-2025-report", "title": "Ethical Considerations and Recommendations for the Use of AI by Judges and Judicial Staff (2025 Annual Report)", "body": "New York State Unified Court System Advisory Committee on AI and the Courts", "bodyShort": "NY UCS AI Advisory Committee", "jurisdiction": "us-state", "region": "New York", "type": "report", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "competence", "confidentiality", "verification" ], "effectiveDate": "2025-12-01T00:00:00.000Z", "versions": [ { "version": "2025 Annual Report", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.nycourts.gov/LegacyPDFS/publications/pdfs/AI%20Annual%20Report%202025.pdf" } ], "summary": "The December 2025 annual report of the New York court system's Advisory Committee on AI and the Courts, containing principles-based ethical considerations and recommendations for judges and judicial staff using AI. It supplements the October 2025 interim policy, addressing impartiality, confidentiality, competence and the need for human judgment in every judicial decision.", "keyProvisions": [ "Judges must ensure AI use is consistent with the Rules Governing Judicial Conduct.", "AI may assist with research and drafting but must not replace judicial reasoning.", "Recommends ongoing training and periodic review of approved tools.", "Reinforces the interim policy's ban on confidential data in public tools." ], "relatedIncidents": [], "sources": [ { "label": "NY Courts — AI Annual Report 2025 (PDF)", "url": "https://www.nycourts.gov/LegacyPDFS/publications/pdfs/AI%20Annual%20Report%202025.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-ma-sjc-interim-genai-guidelines", "url": "https://safelegalai.com/regulation/us-state/us-ma-sjc-interim-genai-guidelines", "title": "Massachusetts Supreme Judicial Court: Interim Guidelines for Use of Generative AI", "body": "Massachusetts Supreme Judicial Court", "bodyShort": "Massachusetts SJC", "jurisdiction": "us-state", "region": "Massachusetts", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality" ], "versions": [ { "version": "Interim", "date": "2025-12-01T00:00:00.000Z", "url": "https://www.mass.gov/guidance/interim-guidelines-for-use-of-generative-ai", "changes": "Announced in the December 2025 State of the Judiciary address; the source gives a month only (announced December 2025), so the day is recorded as the 1st" } ], "summary": "Massachusetts' first judicial AI instrument. The SJC's interim guidelines let court personnel use generative AI for administrative work using only public information, bar tools that retain or train on entered information, and deliberately leave attorney conduct untouched pending possible later rules.", "keyProvisions": [ "Apply to court personnel only: judges, clerks, registers, recorders, employees, law clerks, interns and contractors", "Permit generative AI for administrative tasks using only public information", "Prohibit tools that retain or train on entered information", "Do not govern attorney conduct and do not require attorney AI disclosure in filings", "Described by Chief Justice Budd as a modest first step; subject to revision" ], "relatedIncidents": [], "sources": [ { "label": "SJC Interim Guidelines for Use of Generative AI", "url": "https://www.mass.gov/guidance/interim-guidelines-for-use-of-generative-ai" } ], "country": "US", "categories": [ "judicial-use", "court-own-use", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "uk-bar-council-generative-ai-guidance", "url": "https://safelegalai.com/regulation/uk/uk-bar-council-generative-ai-guidance", "title": "Considerations when using ChatGPT and generative AI software based on large language models", "body": "Bar Council of England and Wales", "bodyShort": "Bar Council", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "barristers" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "effectiveDate": "2024-01-30T00:00:00.000Z", "versions": [ { "version": "November 2025 update", "date": "2025-11-25T00:00:00.000Z", "url": "https://www.barcouncil.org.uk/resource/updated-guidance-on-generative-ai-for-the-bar.html", "changes": "Updated after Ayinde v Haringey to strengthen the verification expectation and address hallucinated authorities, confidentiality of instructions and the Core Duties." }, { "version": "Original", "date": "2024-01-30T00:00:00.000Z", "url": "https://www.barcouncilethics.co.uk/documents/considerations-when-using-chatgpt-and-generative-ai-software-based-on-large-language-models/" } ], "summary": "The Bar Council's ethics guidance for barristers in England and Wales on generative AI, first issued January 2024 and updated November 2025. Barristers may use these tools but must verify all output, avoid disclosing confidential or privileged material to public systems, and remain responsible for their work under the Core Duties.", "keyProvisions": [ "Barristers must independently verify any legal proposition or authority produced by generative AI.", "Client and privileged information should not be entered into public tools without safeguards.", "Barristers should understand the tool's limitations, including hallucination and bias.", "Use of AI does not alter the barrister's personal responsibility under the BSB Handbook." ], "relatedIncidents": [ "ayinde-v-haringey" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "Bar Council — Updated guidance on generative AI for the Bar", "url": "https://www.barcouncil.org.uk/resource/updated-guidance-on-generative-ai-for-the-bar.html" }, { "label": "Bar Council Ethics Hub — original document", "url": "https://www.barcouncilethics.co.uk/documents/considerations-when-using-chatgpt-and-generative-ai-software-based-on-large-language-models/" } ], "country": "GB", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "tr-kvkk-generative-ai-guide", "url": "https://safelegalai.com/regulation/other/tr-kvkk-generative-ai-guide", "title": "Generative AI and the Protection of Personal Data Guide, in 15 Questions (Üretken Yapay Zekâ ve Kişisel Verilerin Korunması Rehberi)", "body": "Personal Data Protection Authority [Kişisel Verileri Koruma Kurumu, KVKK]", "bodyShort": "KVKK", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "consent", "confidentiality", "record-keeping", "competence" ], "versions": [ { "version": "Publication No. 113", "date": "2025-11-24T00:00:00.000Z", "url": "https://www.kvkk.gov.tr/SharedFolderServer/CMSFiles/MTY5MjNmNmIwZWY3YTE.pdf", "changes": "Supplements KVKK's 2021 Recommendations on the Protection of Personal Data in the Field of Artificial Intelligence." } ], "summary": "A 64-page KVKK guide, published November 2025, answering fifteen questions on how Law No. 6698 applies to generative AI. It covers the model lifecycle, controller/processor allocation, the general processing principles, lawful bases, cross-border transfers, transparency, data-subject rights, security, and practical guidance for everyday users and for children's use of AI tools.", "keyProvisions": [ "Sets out how the general principles of Law No. 6698 apply at each stage of a generative AI model's lifecycle.", "Addresses how controller and processor roles are allocated across developers, deployers and users of generative AI.", "Explains the lawful bases available for processing personal data in generative AI systems.", "Covers the regime for transferring personal data abroad in generative AI contexts.", "Sets expectations on transparency, exercise of data-subject rights and data security in generative AI systems." ], "relatedIncidents": [], "sources": [ { "label": "KVKK announcement (24 Nov 2025)", "url": "https://www.kvkk.gov.tr/Icerik/8547/uretken-yapay-zeka-ve-kisisel-verilerin-korunmasi-rehberi-15-soruda" }, { "label": "Guide (official PDF)", "url": "https://www.kvkk.gov.tr/SharedFolderServer/CMSFiles/MTY5MjNmNmIwZWY3YTE.pdf" }, { "label": "KVKK Recommendations on Personal Data Protection in the Field of AI (2021)", "url": "https://www.kvkk.gov.tr/Icerik/7048/Yapay-Zeka-Alaninda-Kisisel-Verilerin-Korunmasina-Dair-Tavsiyeler" } ], "country": "TR", "categories": [ "data-protection", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-va-legal-ethics-opinion-1901", "url": "https://safelegalai.com/regulation/us-state/us-va-legal-ethics-opinion-1901", "title": "Legal Ethics Opinion 1901: Reasonable Fees and the Use of Generative Artificial Intelligence", "body": "Virginia State Bar (approved by the Supreme Court of Virginia)", "bodyShort": "Virginia State Bar", "jurisdiction": "us-state", "region": "Virginia", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "disclosure" ], "effectiveDate": "2025-11-24T00:00:00.000Z", "versions": [ { "version": "Approved", "date": "2025-11-24T00:00:00.000Z", "url": "https://vsb.org/common/Uploaded%20files/LEOs/1901.pdf", "changes": "Approved by the Supreme Court of Virginia and effective immediately; VSB Council approved the proposed opinion on 12 June 2025" } ], "summary": "LEO 1901 addresses how Virginia lawyers may bill for work assisted by generative AI. It concludes that value-based and flat fees may reflect efficiency gains, the skill of effectively incorporating the technology, and the value of the output, and remain reasonable under Rule 1.5(a); a lawyer need not reduce a flat or value-based fee solely because AI accelerated the work.", "keyProvisions": [ "Value-based fees may reflect efficiency gains and the specialised skill of effectively incorporating AI, and remain reasonable under Rule 1.5(a)", "A lawyer is not ethically required to reduce a flat or value-based fee solely because generative AI accelerated the work", "The 'skill requisite to perform the legal service properly' may increase when using AI", "Rule 1.5(b) may require additional explanation of why the lawyer's experience or technical skill contributes value even where time spent has fallen" ], "relatedIncidents": [], "sources": [ { "label": "LEO 1901 (Virginia State Bar official PDF)", "url": "https://vsb.org/common/Uploaded%20files/LEOs/1901.pdf" }, { "label": "LEO 1901 (Supreme Court of Virginia copy)", "url": "https://www.vacourts.gov/static/courts/scv/amendments/leo_1901.pdf" }, { "label": "VSB news release on proposed LEO 1901", "url": "https://vsb.org/Site/Site/news/rules-news/20250327-prop-1901-fees-ai.aspx" } ], "country": "US", "categories": [ "fees-billing", "bar-guidance", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "si-ziudhpui-ai-act-implementation", "url": "https://safelegalai.com/regulation/other/si-ziudhpui-ai-act-implementation", "title": "Act implementing Regulation (EU) 2024/1689 on artificial intelligence (ZIUDHPUI)", "body": "National Assembly of the Republic of Slovenia (Drzavni zbor)", "bodyShort": "Drzavni zbor", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2025-11-21T00:00:00.000Z", "versions": [ { "version": "Adopted 23 October 2025; Uradni list RS 85/25 of 6 November 2025; in force 21 November 2025", "date": "2025-11-21T00:00:00.000Z", "url": "https://www.uradni-list.si/novice/pogled/kdo-bo-nadziral-umetno-inteligenco-v-sloveniji-osnutek-zakona-v-javni-razpravi-/kdo-bo-nadziral-umetno-inteligenco-v-sloveniji-osnutek-zakona-v-javni-razpravi-", "changes": "Original; official gazette text URL not obtained, publisher news page cited." } ], "summary": "Slovenia's AI Act implementing statute designating five market-surveillance authorities: the Information Commissioner for prohibited practices and high-risk systems in biometrics, criminal justice, migration and democratic processes; AKOS as single contact point and sandbox operator; the Bank of Slovenia; the Insurance Supervision Agency; and the Market Inspectorate. It also sets procedures, offences and fines.", "keyProvisions": [ "The Information Commissioner supervises justice-related high-risk AI systems.", "AKOS is single point of contact and regulatory-sandbox administrator.", "Notifying ministries and the Inspectorate for the Information Society oversee notified bodies.", "Offences and fines for AI Act breaches." ], "relatedIncidents": [], "sources": [ { "label": "Uradni list news on the draft act", "url": "https://www.uradni-list.si/novice/pogled/kdo-bo-nadziral-umetno-inteligenco-v-sloveniji-osnutek-zakona-v-javni-razpravi-/kdo-bo-nadziral-umetno-inteligenco-v-sloveniji-osnutek-zakona-v-javni-razpravi-" }, { "label": "Jadek & Pensa summary of ZIUDHPUI (secondary)", "url": "https://www.jadek-pensa.si/kaj-prinasa-zakon-o-izvajanju-akta-o-umetni-inteligenci/" } ], "country": "SI", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "au-wa-supreme-court-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/australia/au-wa-supreme-court-generative-ai-guidelines", "title": "Guidelines for the use of generative AI", "body": "Supreme Court of Western Australia", "bodyShort": "Supreme Court of WA", "jurisdiction": "australia", "region": "Western Australia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence" ], "effectiveDate": "2025-11-20T00:00:00.000Z", "versions": [ { "version": "Published November 2025", "date": "2025-11-20T00:00:00.000Z", "url": "https://www.supremecourt.wa.gov.au/_files/Guidelines%20for%20the%20Use%20of%20Generative%20Artificial%20Intelligence.pdf", "changes": "First issue, following the court's early-2025 consultation note. The document carries no date on its face; 2025-11-20 is taken from the PDF's creation metadata and should be re-confirmed against the court before it is treated as authoritative." } ], "summary": "Principles-based guidelines of the Supreme Court of Western Australia for lawyers, self-represented litigants and expert witnesses. AI-generated content relied on in proceedings must be verified by a human who takes legal responsibility; confidential, suppressed or privileged material must not be entered into public tools; and AI use must be disclosed when the court directs.", "keyProvisions": [ "AI-generated content must be verified by a human who takes legal responsibility for the document; that a document was AI-assisted is no answer to errors or omissions.", "Confidential, suppressed or legally privileged information must not be entered into public or unsecured generative AI tools.", "Affidavits and witness statements must be drafted in the witness's own words; AI content must not substitute the witness's voice.", "Expert reports must comply with the Harmonised Expert Witness Code of Conduct (Practice Direction 4.5.2.1); experts should consider what disclosure of AI use is needed.", "Generative AI must not be used to mislead the court or other parties about the work undertaken; use should be disclosed when directed or otherwise appropriate." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of WA — Guidelines for the use of generative AI (official PDF)", "url": "https://www.supremecourt.wa.gov.au/_files/Guidelines%20for%20the%20Use%20of%20Generative%20Artificial%20Intelligence.pdf" } ], "country": "AU", "categories": [ "verification-duty", "litigant-in-person", "confidentiality-client-data", "evidence-admissibility", "disclosure-filings", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cl-corte-suprema-acta-244-2025-codigo-etica-judicial", "url": "https://safelegalai.com/regulation/other/cl-corte-suprema-acta-244-2025-codigo-etica-judicial", "title": "Acta 244-2025: Judicial Ethics Code (article 17 on responsible use of technology and AI)", "body": "Supreme Court of Chile (Corte Suprema)", "bodyShort": "Corte Suprema", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "prohibition", "confidentiality", "supervision" ], "effectiveDate": "2025-11-19T00:00:00.000Z", "versions": [ { "version": "Approved at the XXVIII Jornadas de Reflexion (23-24 October 2025); published in the Diario Oficial", "date": "2025-11-19T00:00:00.000Z", "url": "https://www.bcn.cl/leychile/Navegar?idNorma=1218674", "changes": "First Judicial Ethics Code and first AI rule for Chilean judges." } ], "summary": "Chile's first Judicial Ethics Code, adopted by auto acordado and published on 19 November 2025. Article 17 permits AI as a support tool but declares judicial decisions always human and non-delegable, requires judges to verify any AI result independently, forbids automatic incorporation of AI text into rulings, and requires control over data shared with external platforms.", "keyProvisions": [ "Judicial decisions are human and non-delegable; no delegation of evidence assessment or dispute resolution to AI.", "Judges must independently verify AI-generated references.", "No automatic or unreflective incorporation of AI text into filings or judgments.", "Automated transcription and translation only as preliminary support, reviewed and certified by a person.", "Control of data shared with external platforms; confidentiality in virtual hearings." ], "relatedIncidents": [], "sources": [ { "label": "Ley Chile (BCN) — Acta 244-2025", "url": "https://www.bcn.cl/leychile/Navegar?idNorma=1218674" }, { "label": "Direccion de Estudios, Poder Judicial", "url": "https://direcciondeestudios.pjud.cl/corte-suprema-aprueba-su-primer-codigo-de-etica-judicial-un-hito-a-la-excelencia-y-la-confianza-ciudadana-en-el-poder-judicial" } ], "country": "CL", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "law-society-ireland-generative-ai-guidance", "url": "https://safelegalai.com/regulation/ireland/law-society-ireland-generative-ai-guidance", "title": "Use of Generative AI: Guidance for the Legal Profession", "body": "Law Society of Ireland", "bodyShort": "Law Society of Ireland", "jurisdiction": "ireland", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "confidentiality", "verification", "disclosure", "competence" ], "effectiveDate": "2025-11-12T00:00:00.000Z", "versions": [ { "version": "Version 4 (12 November 2025)", "date": "2025-11-12T00:00:00.000Z", "url": "https://www.lawsociety.ie/globalassets/documents/cybersecurity/2025-12-generative-ai-guidelines-v4.pdf" } ], "summary": "The Law Society of Ireland's generative AI guidelines for solicitors, at version 4 dated 12 November 2025. They identify suitable and unsuitable uses, require client confidentiality to be protected when information is entered into tools, call for transparency with clients, require verification of output, and address the AI literacy duty in Article 4 of the EU AI Act.", "keyProvisions": [ "Lists suitable and unsuitable uses of generative AI in solicitors' practice.", "Client confidential information must not be entered into tools without adequate safeguards.", "Solicitors should be transparent with clients about material AI use.", "Firms must address the EU AI Act Article 4 AI literacy obligation for staff.", "All output must be verified; the solicitor remains responsible." ], "relatedIncidents": [ "guerin-v-odoherty" ], "article": "ireland-ai-guidance-history", "sources": [ { "label": "Law Society of Ireland — generative AI guidelines v4 (PDF)", "url": "https://www.lawsociety.ie/globalassets/documents/cybersecurity/2025-12-generative-ai-guidelines-v4.pdf" }, { "label": "Law Society of Ireland — generative AI guidance page", "url": "https://www.lawsociety.ie/artificial-intelligence-ai/generative-ai-guidance/" } ], "country": "IE", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "client-consent-transparency", "competence-duty", "supervision", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ph-sc-2025-code-judicial-conduct-accountability", "url": "https://safelegalai.com/regulation/other/ph-sc-2025-code-judicial-conduct-accountability", "title": "A.M. No. 25-04-04-SC: The 2025 Code of Judicial Conduct and Accountability", "body": "Supreme Court of the Philippines (En Banc)", "bodyShort": "Supreme Court of the Philippines", "jurisdiction": "other", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "verification", "prohibition", "competence" ], "effectiveDate": "2025-11-11T00:00:00.000Z", "versions": [ { "version": "Promulgated", "date": "2025-11-11T00:00:00.000Z", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/100836" } ], "summary": "The Philippines' new judicial conduct code, promulgated 11 November 2025, is one of the first anywhere to write artificial-intelligence limits into a code of judicial ethics rather than into guidance. Canon VI addresses competence, diligence and the ethical use of technology and artificial intelligence by judges.", "keyProvisions": [ "Canon VI, Section 3: a judge shall personally write all decisions and resolutions.", "Canon VI, Section 5: a judge may use AI tools only for research of relevant legal authorities, and must always verify the sources before citing them.", "Canon VI, Section 6: ethical use of technology and artificial intelligence." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court E-Library — A.M. No. 25-04-04-SC full text", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/100836" } ], "country": "PH", "categories": [ "judicial-use", "verification-duty", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "co-corte-suprema-stc17832-2025", "url": "https://safelegalai.com/regulation/other/co-corte-suprema-stc17832-2025", "title": "STC17832-2025: annulment of a judgment based on non-existent jurisprudential citations", "body": "Supreme Court of Justice, Civil Cassation Chamber (Corte Suprema de Justicia, Sala de Casacion Civil)", "bodyShort": "CSJ Sala Civil", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "judges", "lawyers" ], "requirements": [ "verification" ], "effectiveDate": "2025-11-05T00:00:00.000Z", "versions": [ { "version": "Judgment, MP Adriana Lopez Martinez", "date": "2025-11-05T00:00:00.000Z", "url": "https://www.infobae.com/colombia/2025/11/12/historico-fallo-de-la-corte-suprema-sobre-el-uso-de-la-ia-en-colombia-por-primera-vez-anulo-una-decision-judicial-apoyada-en-citas-falsas/", "changes": "Initial decision; reported by Infobae, official text not obtained." } ], "summary": "The Supreme Court annulled a Sincelejo Superior Tribunal ruling that quoted passages from two cassation decisions which do not exist, treating reliance on unverifiable sources as a via de hecho breaching due process, and ordered the Judicial Council and the Rodrigo Lara Bonilla Judicial School to disseminate the decision to all judges.", "keyProvisions": [ "Judicial decisions must rest on verifiable sources; fabricated citations vitiate the ruling.", "Judges owe high diligence in verifying cited jurisprudence.", "Dissemination ordered through the Judicial School.", "A companion ruling fined a lawyer 15 SMLMV under CGP arts 80-81 for ten AI-generated fake rulings." ], "relatedIncidents": [], "sources": [ { "label": "Infobae report", "url": "https://www.infobae.com/colombia/2025/11/12/historico-fallo-de-la-corte-suprema-sobre-el-uso-de-la-ia-en-colombia-por-primera-vez-anulo-una-decision-judicial-apoyada-en-citas-falsas/" }, { "label": "El Tiempo on the lawyer's fine", "url": "https://www.eltiempo.com/justicia/cortes/la-inedita-sancion-de-la-corte-suprema-a-un-abogado-que-uso-la-inteligencia-artificial-para-redactar-un-recurso-de-revision-3532886" } ], "country": "CO", "categories": [ "sanctions-record", "verification-duty", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "in-meity-india-ai-governance-guidelines-2025", "url": "https://safelegalai.com/regulation/india/in-meity-india-ai-governance-guidelines-2025", "title": "India AI Governance Guidelines", "body": "Ministry of Electronics and Information Technology (IndiaAI Mission)", "bodyShort": "MeitY", "jurisdiction": "india", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "permissive" ], "effectiveDate": "2025-11-05T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-11-05T00:00:00.000Z", "url": "https://www.newsonair.gov.in/meity-unveils-india-ai-governance-guidelines-to-promote-safe-and-responsible-ai-adoption", "changes": "Released 5 November 2025; official PDF hosted on indiaai.gov.in was not retrievable." } ], "summary": "Non-binding national AI governance framework released by MeitY on 5 November 2025 under the IndiaAI Mission: seven guiding principles, recommendations across six governance pillars, and a sectoral approach leaving domain rules to sectoral regulators. It contains no risk classification for justice or legal uses and creates no new statute.", "keyProvisions": [ "Seven guiding principles for responsible AI: trust, people first, innovation over restraint, fairness, accountability, understandable by design, safety.", "Sectoral regulation preferred over a horizontal AI statute.", "Action plan across short, medium and long-term timelines." ], "relatedIncidents": [], "sources": [ { "label": "News on AIR (official government news service)", "url": "https://www.newsonair.gov.in/meity-unveils-india-ai-governance-guidelines-to-promote-safe-and-responsible-ai-adoption" }, { "label": "Chambers and Partners analysis", "url": "https://chambers.com/articles/meity-unveils-india-s-approach-towards-regulating-artificial-intelligence" } ], "country": "IN", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "es-icam-guia-buenas-practicas-ia-abogacia-2025", "url": "https://safelegalai.com/regulation/other/es-icam-guia-buenas-practicas-ia-abogacia-2025", "title": "ICAM Guide of Good Practice for the use of AI in the legal profession (Guia ICAM IA 2025)", "body": "Madrid Bar Association (Ilustre Colegio de la Abogacia de Madrid)", "bodyShort": "ICAM", "jurisdiction": "other", "region": "Madrid", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence" ], "versions": [ { "version": "October 2025", "date": "2025-10-31T00:00:00.000Z", "url": "https://web.icam.es//wp-content/uploads/2025/10/Guia-ICAM_IA_2025_2.pdf", "changes": "Original (exact day not shown on the page)." } ], "summary": "Madrid Bar guide keeping human professional judgment at the centre of AI-assisted practice: verify outputs, protect client data, inform clients about AI use, bill time transparently, train lawyers, and vet vendors on no-training clauses, EU data residency and Article 28 GDPR agreements. It includes an AI glossary, an EU AI Act compliance timeline and infographics.", "keyProvisions": [ "Validate all AI outputs; models produce plausible but inaccurate content.", "Manage client data carefully under confidentiality and data-protection duties.", "Inform clients about the use of AI tools.", "Bill AI-assisted time transparently.", "Vendor selection criteria and EU AI Act obligations for firms." ], "relatedIncidents": [], "sources": [ { "label": "ICAM — AI and the legal profession", "url": "https://web.icam.es/abogacia-inteligencia-artificial/" }, { "label": "ICAM guide (PDF)", "url": "https://web.icam.es//wp-content/uploads/2025/10/Guia-ICAM_IA_2025_2.pdf" } ], "country": "ES", "categories": [ "bar-guidance", "confidentiality-client-data", "client-consent-transparency", "fees-billing", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "uk-judiciary-ai-guidance", "url": "https://safelegalai.com/regulation/uk/uk-judiciary-ai-guidance", "title": "Artificial Intelligence (AI): Judicial Guidance", "body": "Courts and Tribunals Judiciary of England and Wales", "bodyShort": "UK Judiciary", "jurisdiction": "uk", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "confidentiality", "verification", "competence" ], "effectiveDate": "2023-12-12T00:00:00.000Z", "versions": [ { "version": "October 2025 (v3)", "date": "2025-10-31T00:00:00.000Z", "url": "https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance-october-2025/", "changes": "Replaces the April 2025 version; expands guidance on identifying AI-generated material from litigants and lawyers and on judges' personal responsibility for anything produced with AI assistance." }, { "version": "April 2025 (v2)", "date": "2025-04-14T00:00:00.000Z", "url": "https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance/", "changes": "Extended to all judicial office holders, added a glossary, and addressed hallucinated case law and deepfake evidence." }, { "version": "December 2023 (v1)", "date": "2023-12-12T00:00:00.000Z", "url": "https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance/" } ], "summary": "Guidance from the senior judiciary of England and Wales for judges, magistrates and tribunal members on using AI. Judges remain personally responsible for their work, must not enter confidential information into public tools, and may ask parties and lawyers about their use of AI. The current version is dated 31 October 2025.", "keyProvisions": [ "Judicial office holders are personally responsible for material produced in their name, whether or not AI assisted.", "Confidential or private information must not be entered into public AI chatbots.", "Judges should be alert to fabricated citations and may ask parties and representatives whether AI was used.", "AI may be used for summarising and administrative tasks but not for legal research or analysis that cannot be independently verified." ], "relatedIncidents": [ "ayinde-v-haringey", "harber-v-hmrc", "ms-bangladesh-2025-ukut-305", "father-v-mother-2025-ewhc-2135-fam" ], "article": "uk-judicial-ai-guidance-version-history", "sources": [ { "label": "Judiciary of England and Wales — AI judicial guidance (October 2025)", "url": "https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance-october-2025/" }, { "label": "Judiciary of England and Wales — earlier guidance page (December 2023 and April 2025)", "url": "https://www.judiciary.uk/guidance-and-resources/artificial-intelligence-ai-judicial-guidance/" } ], "country": "GB", "categories": [ "judicial-use", "ai-decision-prohibition", "confidentiality-client-data", "verification-duty", "court-own-use", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-mt-4th-district-rule-3g-generative-ai", "url": "https://safelegalai.com/regulation/us-state/us-mt-4th-district-rule-3g-generative-ai", "title": "Montana Fourth Judicial District Court Rules of Practice, Rule 3.G — Pleading Requirements for Use of Generative Artificial Intelligence", "body": "Montana Fourth Judicial District Court (Missoula and Mineral Counties)", "bodyShort": "MT 4th Judicial District Court", "jurisdiction": "us-state", "region": "Montana", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "certification", "verification" ], "versions": [ { "version": "Rev. October 31, 2025", "date": "2025-10-31T00:00:00.000Z", "url": "https://courts.mt.gov/external/dcourt/dc_rules/4th.pdf", "changes": "Rule 3.G added to the district's Rules of Practice" } ], "summary": "Rule 3.G of the Montana Fourth Judicial District Court's Rules of Practice requires any party who used generative AI in preparing a filed document to disclose that use, name the tool, explain how it was used, and certify that AI-assisted portions have been checked for accuracy. A filing without the certification certifies that no AI was used.", "keyProvisions": [ "Any party using generative AI to prepare a filed pleading or document must disclose the use, the specific tool and how it was used", "The party must certify they have checked the accuracy of any AI-drafted or AI-assisted portion, including factual and procedural background, citations and legal authority", "A filing without the certification is presumed to certify that no part was prepared using generative AI", "Responsibility for content rests with the attorney or pro se litigant under Montana Rule of Civil Procedure 11 and the rules of professional conduct", "Suspected violations may result in Rule 11 sanctions including dismissal of the party's case, document or pleading without prejudice" ], "relatedIncidents": [], "sources": [ { "label": "MT 4th Judicial District Rules of Practice (official PDF)", "url": "https://courts.mt.gov/external/dcourt/dc_rules/4th.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-md-mezu-v-mezu", "url": "https://safelegalai.com/regulation/us-state/us-md-mezu-v-mezu", "title": "Mezu v. Mezu, No. 361, Sept. Term 2025 (Appellate Court of Maryland, reported, 29 Oct 2025)", "body": "Appellate Court of Maryland", "bodyShort": "Appellate Court of Maryland", "jurisdiction": "us-state", "region": "Maryland", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "supervision" ], "versions": [ { "version": "Reported opinion", "date": "2025-10-29T00:00:00.000Z", "url": "https://www.mdcourts.gov/data/opinions/cosa/2025/0361s25.pdf", "changes": "Reported opinion issued" } ], "summary": "Maryland's leading AI-hallucination decision. The Appellate Court of Maryland published Mezu as a warning to the bar, holding that failing to read AI-generated citations before filing can produce briefs citing multiple fictitious cases, and referred the signing attorney to the Attorney Grievance Commission. Delegation to a law clerk was no defence.", "keyProvisions": [ "The court reported the case to address the recurring problem of AI-drafted briefs containing incorrect, inaccurate or fictitious citations", "The brief contained citations to multiple fictitious cases, misquoted passages, and citations that did not support the propositions cited", "Failure to read AI-generated citations before filing is the operative breach", "Based on the nature and severity of the conduct the court referred the attorney to the Attorney Grievance Commission", "Delegation of research to a law clerk did not excuse the signing attorney" ], "relatedIncidents": [], "sources": [ { "label": "Mezu v. Mezu (PDF, official)", "url": "https://www.mdcourts.gov/data/opinions/cosa/2025/0361s25.pdf" }, { "label": "MSBA announcement", "url": "https://www.msba.org/site/site/content/News-and-Publications/News/General-News/Maryland_Lawyer_Referred_to_Attorney_Grievance_Commission_for_Citing_Fake%20AI_Cases.aspx" } ], "country": "US", "categories": [ "sanctions-record", "verification-duty", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ph-sc-2025-code-conduct-court-officials-personnel", "url": "https://safelegalai.com/regulation/other/ph-sc-2025-code-conduct-court-officials-personnel", "title": "A.M. No. 25-06-11-SC: The 2025 Code of Conduct and Accountability for Court Officials and Personnel", "body": "Supreme Court of the Philippines (En Banc)", "bodyShort": "Supreme Court of the Philippines", "jurisdiction": "other", "type": "court-rule", "status": "in-force", "appliesTo": [ "court-staff" ], "requirements": [ "verification", "competence" ], "effectiveDate": "2025-10-28T00:00:00.000Z", "versions": [ { "version": "Promulgated", "date": "2025-10-28T00:00:00.000Z", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/100842" } ], "summary": "The companion conduct code for Philippine court officials and personnel, promulgated on 28 October 2025. Canon III, Section 6 mirrors the judicial code's AI provision, confining AI use to research of legal authorities and requiring the verification of sources before they are cited in any court document.", "keyProvisions": [ "Canon III, Section 6 restricts AI use to research of legal authorities and requires verification of sources before citation.", "Applies to all court officials and personnel rather than to judges, who are covered by A.M. No. 25-04-04-SC." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court E-Library — A.M. No. 25-06-11-SC full text", "url": "https://elibrary.judiciary.gov.ph/thebookshelf/showdocs/11/100842" } ], "country": "PH", "categories": [ "judicial-use", "verification-duty", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ut-judicial-council-interim-genai-rules", "url": "https://safelegalai.com/regulation/us-state/us-ut-judicial-council-interim-genai-rules", "title": "Utah Judicial Council Interim Rules on the Use of Generative AI", "body": "Utah Judicial Council", "bodyShort": "Utah Judicial Council", "jurisdiction": "us-state", "region": "Utah", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "permissive", "confidentiality" ], "versions": [ { "version": "Interim rules", "date": "2025-10-27T00:00:00.000Z", "url": "https://sites.google.com/utcourts.gov/courts-hr/policy/ai-information", "changes": "Interim rules implemented for court-related work and court-owned devices" } ], "summary": "From 27 October 2025 the Utah Judicial Council operates an allow-list policy for generative AI in court-related work and on court-owned devices. Eight named tools are expressly permitted and any AI tool not expressly permitted is prohibited, giving Utah one of the most restrictive US state judiciary AI postures.", "keyProvisions": [ "Applies to generative AI used for court-related work or on court-owned devices from 27 October 2025", "Eight tools are expressly permitted: Adobe Firefly, ChatGPT, Claude.ai (Beta), Gemini Pro, Microsoft Copilot, NotebookLM, Webex Cisco AI Assistant and Westlaw", "Any AI tool not expressly permitted in the interim rules is prohibited" ], "relatedIncidents": [], "sources": [ { "label": "Utah Judiciary HR, AI Information (official)", "url": "https://sites.google.com/utcourts.gov/courts-hr/policy/ai-information" } ], "country": "US", "categories": [ "judicial-use", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cl-boletin-16821-19-ai-systems-bill", "url": "https://safelegalai.com/regulation/other/cl-boletin-16821-19-ai-systems-bill", "title": "Bill regulating artificial intelligence systems (Boletin 16821-19, merged with 15869-19)", "body": "National Congress of Chile (Congreso Nacional)", "bodyShort": "Congreso", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Second constitutional stage, Senate Future Challenges Committee", "date": "2025-10-24T00:00:00.000Z", "url": "https://www.senado.cl/comunicaciones/noticias/proyecto-que-regula-sistemas-de-inteligencia-artificial-sera-estudiado-por", "changes": "Passed the Chamber of Deputies; under Senate committee study." }, { "version": "Executive message introduced", "date": "2024-05-07T00:00:00.000Z", "url": "https://www.camara.cl/legislacion/proyectosdeley/tramitacion.aspx?prmID=17429&prmBOLETIN=16821-19" } ], "summary": "Government bill classifying AI systems as unacceptable, high, limited or no-evident risk, with obligations for high-risk deployers, a Technical Advisory Council under the Ministry of Science, and enforcement through the future Personal Data Protection Agency. Approved by the Chamber of Deputies and before the Senate's Future Challenges Committee since October 2025.", "keyProvisions": [ "Four risk tiers modelled on the EU AI Act.", "Technical Advisory Council on AI proposing high-risk and limited-risk lists.", "Obligations of transparency and human oversight for high-risk systems.", "Whether justice uses are listed as high risk was not verified." ], "relatedIncidents": [], "sources": [ { "label": "Camara de Diputados — tramitacion Boletin 16821-19", "url": "https://www.camara.cl/legislacion/proyectosdeley/tramitacion.aspx?prmID=17429&prmBOLETIN=16821-19" }, { "label": "Senado news", "url": "https://www.senado.cl/comunicaciones/noticias/proyecto-que-regula-sistemas-de-inteligencia-artificial-sera-estudiado-por" } ], "country": "CL", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-il-ardc-implementing-generative-ai-guide", "url": "https://safelegalai.com/regulation/us-state/us-il-ardc-implementing-generative-ai-guide", "title": "ARDC: Implementing Generative AI for Illinois Attorneys", "body": "Illinois Attorney Registration and Disciplinary Commission", "bodyShort": "Illinois ARDC", "jurisdiction": "us-state", "region": "Illinois", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "verification", "supervision", "consent" ], "versions": [ { "version": "Original", "date": "2025-10-24T00:00:00.000Z", "url": "https://iardc.org/Files/Implementing-AI-Guide/?page=1", "changes": "Initial publication" } ], "summary": "The Illinois ARDC's practical companion to the Supreme Court's AI Policy. It is expressly nonbinding and subordinate to the Policy, and supplies an explanation of generative AI, a tool-selection and documentation framework, and a Practice Resource Kit with sample notices, firm policies, a client consent form and a terms-of-use checklist.", "keyProvisions": [ "Nonbinding companion to the Illinois Supreme Court Policy on Artificial Intelligence; the Policy controls in the event of inconsistency", "Explains how modern generative AI works and why that matters for confidentiality and security", "Provides a framework to choose and document appropriate AI tools", "Practice Resource Kit: sample client notices, firm AI policies, client consent form and terms-of-use checklist" ], "relatedIncidents": [], "sources": [ { "label": "ARDC Artificial Intelligence resources page", "url": "https://www.iardc.org/EducationAndOutreach/ArtificialIntelligence" }, { "label": "Implementing Generative AI guide", "url": "https://iardc.org/Files/Implementing-AI-Guide/?page=1" } ], "country": "US", "categories": [ "bar-guidance", "confidentiality-client-data", "competence-duty", "verification-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-sd-ohio-graham-genai-standing-order", "url": "https://safelegalai.com/regulation/us-federal/us-sd-ohio-graham-genai-standing-order", "title": "Standing Order on the Use of Generative Artificial Intelligence (Judge James L. Graham)", "body": "United States District Court for the Southern District of Ohio, Judge James L. Graham", "bodyShort": "S.D. Ohio (Graham)", "jurisdiction": "us-federal", "region": "Ohio", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "disclosure", "certification", "verification" ], "effectiveDate": "2025-10-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-10-21T00:00:00.000Z", "url": "https://www.ohsd.uscourts.gov/sites/ohsd/files/10-21-25%20JLG%20Standing%20Order%20on%20the%20Use%20of%20Generative%20AI.pdf" } ], "summary": "A standing order issued 21 October 2025 by Judge James Graham of the Southern District of Ohio. Any filing prepared with generative AI must be accompanied by a declaration disclosing that AI was used, identifying the platform, describing which content was AI-generated, and confirming that a human verified every citation and factual assertion before filing.", "keyProvisions": [ "Declaration required with any filing that used generative AI.", "Declaration must name the platform and describe what content was AI-generated.", "Filer must confirm human verification of all citations, quotations and facts.", "Failure to comply may result in the filing being struck and sanctions." ], "relatedIncidents": [], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "S.D. Ohio — Judge Graham standing order on generative AI (PDF)", "url": "https://www.ohsd.uscourts.gov/sites/ohsd/files/10-21-25%20JLG%20Standing%20Order%20on%20the%20Use%20of%20Generative%20AI.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ar-csjn-acordada-10-2025-ai-summaries", "url": "https://safelegalai.com/regulation/other/ar-csjn-acordada-10-2025-ai-summaries", "title": "Acordada 10/2025: AI-generated plain-language summaries of judgments (LEX100)", "body": "Supreme Court of Justice of the Nation (Corte Suprema de Justicia de la Nacion)", "bodyShort": "CSJN", "jurisdiction": "other", "type": "court-rule", "status": "in-force", "appliesTo": [ "court-staff" ], "requirements": [ "permissive" ], "effectiveDate": "2025-10-15T00:00:00.000Z", "versions": [ { "version": "Implementation announced", "date": "2025-10-15T00:00:00.000Z", "url": "https://www.csjn.gov.ar/novedades/detalle/11349", "changes": "Automatic AI summaries for all judgments published via LEX100." } ], "summary": "Under Acordada 10/2025 the Argentine Supreme Court implemented automatic generation of plain-language summaries by AI for every judgment published from the LEX100 system used by federal and national courts, with more than 110,000 summaries produced by October 2025. The Court also uses AI to classify incoming appeals. Human-review rules were not stated.", "keyProvisions": [ "AI summaries in clear language for all LEX100 judgments.", "Applies across federal and national courts.", "AI classification of appeals filed with the Court.", "Internal AI assistant for jurisprudence and superintendence consultation in development." ], "relatedIncidents": [], "sources": [ { "label": "CSJN news", "url": "https://www.csjn.gov.ar/novedades/detalle/11349" } ], "country": "AR", "categories": [ "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "it-cnf-informativa-ia-art-13", "url": "https://safelegalai.com/regulation/other/it-cnf-informativa-ia-art-13", "title": "CNF communication to local bars with model client notice on AI use (Art. 13 Law 132/2025)", "body": "National Bar Council (Consiglio Nazionale Forense)", "bodyShort": "CNF", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "disclosure" ], "effectiveDate": "2025-10-13T00:00:00.000Z", "versions": [ { "version": "Prot. 0007816, 13 October 2025", "date": "2025-10-13T00:00:00.000Z", "url": "https://www.consiglionazionaleforense.it/documents/20182/0/Comunicazione+ai+COA+-+Schema+informativa+IA.pdf" } ], "summary": "Letter from CNF President Francesco Greco to all local bar presidents noting that Article 13 of Law 132/2025 obliges professionals to give clients an informativa on the AI systems they use, and circulating a CNF model text lawyers may adopt in dealings with clients and assisted parties, to be disseminated to all members.", "keyProvisions": [ "Confirms that Art. 13 of Law 132/2025 creates a duty to inform clients about AI use.", "Provides an optional model informativa for lawyer-client communications.", "Requests maximum dissemination to all registered lawyers through the local bars." ], "relatedIncidents": [], "sources": [ { "label": "CNF communication and model informativa (PDF)", "url": "https://www.consiglionazionaleforense.it/documents/20182/0/Comunicazione+ai+COA+-+Schema+informativa+IA.pdf" } ], "country": "IT", "categories": [ "client-consent-transparency", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "mt-artificial-intelligence-regulations-sl-591-05", "url": "https://safelegalai.com/regulation/other/mt-artificial-intelligence-regulations-sl-591-05", "title": "Artificial Intelligence Regulations, 2025 (Subsidiary Legislation 591.05)", "body": "Government of Malta under the Malta Digital Innovation Authority Act (Cap. 591)", "bodyShort": "Malta (MDIA Act)", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2025-10-10T00:00:00.000Z", "versions": [ { "version": "In force 10 October 2025; most operative provisions from 2 August 2026", "date": "2025-10-10T00:00:00.000Z", "url": "https://legislation.mt/eli/sl/591.5/eng", "changes": "Original; Legal Notice number not captured (landing page opened, PDF not readable)." } ], "summary": "Maltese subsidiary legislation giving effect to the EU AI Act's institutional requirements: the Malta Digital Innovation Authority is market-surveillance authority, notifying authority, single point of contact and operator of the national AI sandbox, with appeals to the Administrative Review Tribunal. A companion instrument allocates high-risk and prohibited-practice responsibilities to the Information and Data Protection Commissioner.", "keyProvisions": [ "MDIA designated as primary national competent and market-surveillance authority.", "National AI regulatory sandbox administered by the MDIA.", "Magistrate authorisation required for post-remote biometric identification by law enforcement." ], "relatedIncidents": [], "sources": [ { "label": "Legizlazzjoni Malta — S.L. 591.05", "url": "https://legislation.mt/eli/sl/591.5/eng" }, { "label": "LexMT Malta AI regulation guide (secondary)", "url": "https://lex.mt/articles/malta-ai-regulation-guide" }, { "label": "INPLP Malta 2025 legal review (secondary)", "url": "https://inplp.com/latest-news/article/malta-at-the-intersection-of-it-law-data-protection-and-artificial-intelligence-a-2025-legal-review/" } ], "country": "MT", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ny-ucs-interim-ai-policy", "url": "https://safelegalai.com/regulation/us-state/ny-ucs-interim-ai-policy", "title": "Interim Policy on the Use of Artificial Intelligence", "body": "New York State Unified Court System", "bodyShort": "NY Unified Court System", "jurisdiction": "us-state", "region": "New York", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "competence", "confidentiality", "verification" ], "effectiveDate": "2025-10-10T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-10-10T00:00:00.000Z", "url": "https://www.nycourts.gov/LegacyPDFS/a.i.-policy.pdf" } ], "summary": "An interim policy adopted by the New York State Unified Court System on 10 October 2025 for judges and non-judicial employees. It permits only approved generative AI tools, requires training, forbids entering confidential or non-public information into public tools, and requires that all output be reviewed and verified by the human user, who remains accountable for the result.", "keyProvisions": [ "Only court-approved AI tools may be used for official work.", "Judges and staff must complete training before using approved tools.", "Confidential, sealed and personal information must not be entered into public AI tools.", "All AI output must be reviewed for accuracy; responsibility stays with the user." ], "relatedIncidents": [], "sources": [ { "label": "NY Courts — Interim AI policy (PDF)", "url": "https://www.nycourts.gov/LegacyPDFS/a.i.-policy.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ccbe-guide-generative-ai-lawyers", "url": "https://safelegalai.com/regulation/eu/ccbe-guide-generative-ai-lawyers", "title": "CCBE Guide on the use of generative AI by lawyers", "body": "Council of Bars and Law Societies of Europe", "bodyShort": "CCBE", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "effectiveDate": "2025-10-02T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-10-02T00:00:00.000Z", "url": "https://www.ccbe.eu/fileadmin/speciality_distribution/public/documents/IT_LAW/ITL_Guides_recommendations/EN_ITL_20251002_CCBE-guide-on-the-use-of-the-use-of-generative-AI-for-lawyers.pdf" } ], "summary": "Guide published 2 October 2025 by the CCBE, which represents European bars and law societies, on how lawyers should use generative AI. It is permissive with cautions: lawyers must understand tool limitations, verify all output, protect professional secrecy and client confidentiality, and consider the EU AI Act's literacy and transparency duties.", "keyProvisions": [ "Lawyers remain fully responsible for advice and documents produced with generative AI.", "Output must be checked for accuracy, including invented citations and misstatements of law.", "Professional secrecy requires care over what is entered into tools and where data is processed.", "Recommends firm-level policies, training and awareness of the AI Act's Article 4 literacy duty." ], "relatedIncidents": [], "article": "does-eu-ai-act-apply-uk-law-firms", "sources": [ { "label": "CCBE — Guide on the use of generative AI by lawyers (PDF)", "url": "https://www.ccbe.eu/fileadmin/speciality_distribution/public/documents/IT_LAW/ITL_Guides_recommendations/EN_ITL_20251002_CCBE-guide-on-the-use-of-the-use-of-generative-AI-for-lawyers.pdf" } ], "country": "EU", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "client-consent-transparency", "data-protection" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ph-house-bill-3462-ai-commission", "url": "https://safelegalai.com/regulation/other/ph-house-bill-3462-ai-commission", "title": "House Bill No. 3462 (20th Congress): creating the Philippine Artificial Intelligence Commission", "body": "House of Representatives of the Philippines (Rep. Salvador A. Pleyto Sr.)", "bodyShort": "House of Representatives", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2025-10-02T00:00:00.000Z", "versions": [ { "version": "Filed", "date": "2025-10-02T00:00:00.000Z", "url": "https://docs.congress.hrep.online/legisdocs/basic_20/HB03462.pdf" } ], "summary": "A pending Philippine House bill, filed 2 October 2025, that would establish a Philippine Artificial Intelligence Commission as the country's national AI regulator. It had not been enacted as at 4 September 2026 and contains no provisions specific to courts, judges or the legal profession.", "keyProvisions": [ "Would create a Philippine Artificial Intelligence Commission as national AI regulator.", "Not enacted as at 4 September 2026." ], "relatedIncidents": [], "sources": [ { "label": "House of Representatives — House Bill No. 3462 (official PDF)", "url": "https://docs.congress.hrep.online/legisdocs/basic_20/HB03462.pdf" } ], "country": "PH", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ar-rio-negro-stj-acordada-15-2024-iagen-protocol", "url": "https://safelegalai.com/regulation/other/ar-rio-negro-stj-acordada-15-2024-iagen-protocol", "title": "Acordada 15/2024 and October 2025 extension: Protocol of Good Practice for Generative AI in the Judiciary of Rio Negro", "body": "Superior Court of Justice of Rio Negro (Superior Tribunal de Justicia de Rio Negro)", "bodyShort": "STJ Rio Negro", "jurisdiction": "other", "region": "Rio Negro", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges", "court-staff", "lawyers" ], "requirements": [ "verification", "supervision", "confidentiality" ], "effectiveDate": "2024-10-01T00:00:00.000Z", "versions": [ { "version": "Acordada extending the protocol to lawyers, experts and other external auxiliaries", "date": "2025-10-01T00:00:00.000Z", "url": "https://justicia.ar/novedades/2029/2025/10/una-nueva-acordada-habilita-sanciones-a-la-abogacia-por-uso-indebido-de-inteligencia-artificial", "changes": "Protocol made mandatory for lawyers; courts may impose corrective sanctions and must refer fabricated citations to bar ethics tribunals." }, { "version": "Acordada 15/2024: protocol for judicial personnel", "date": "2024-10-01T00:00:00.000Z", "url": "https://www.justicia.ar/iniciativas/inteligencia-artificial" } ], "summary": "Rio Negro's judiciary adopted a generative-AI good-practice protocol in October 2024 covering validation of AI content, human control at every stage, bias evaluation and personal-data protection, and in October 2025 extended it to lawyers, experts and other external auxiliaries, authorising corrective sanctions and mandatory referral to bar ethics tribunals for fabricated citations.", "keyProvisions": [ "Validation of AI-produced content and human control at each stage.", "Bias evaluation and personal-data protection.", "Mandatory for lawyers, experts and external auxiliaries from October 2025.", "Courts may apply corrective sanctions and must refer fabricated citations to bar disciplinary bodies." ], "relatedIncidents": [], "sources": [ { "label": "Justicia Rio Negro news (October 2025)", "url": "https://justicia.ar/novedades/2029/2025/10/una-nueva-acordada-habilita-sanciones-a-la-abogacia-por-uso-indebido-de-inteligencia-artificial" }, { "label": "Justicia Rio Negro AI initiatives", "url": "https://www.justicia.ar/iniciativas/inteligencia-artificial" } ], "country": "AR", "categories": [ "judicial-use", "verification-duty", "sanctions-record", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-la-tech-commission-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/us-state/us-la-tech-commission-generative-ai-guidelines", "title": "Louisiana Supreme Court Technology Commission: Generative Artificial Intelligence Guidelines", "body": "Louisiana Supreme Court Technology Commission", "bodyShort": "Louisiana Technology Commission", "jurisdiction": "us-state", "region": "Louisiana", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "lawyers", "court-staff" ], "requirements": [ "verification", "supervision", "confidentiality" ], "versions": [ { "version": "Original", "date": "2025-10-01T00:00:00.000Z", "url": "https://lajudicialcollege.org/LASC%20Tech%20Commission%20Generative%20AI%20Guidelines.pdf", "changes": "Initial publication; the source gives a month only (issued October 2025), so the day is recorded as the 1st" } ], "summary": "The Louisiana Supreme Court's Technology Commission issued advisory generative-AI guidelines in October 2025, chiefly for judges but addressing counsel. They call for active review and verification of AI output analogous to supervising a junior associate and warn against free consumer AI tools in favour of enterprise systems with security protocols.", "keyProvisions": [ "Advisory only, primarily directed at judges but referencing obligations of counsel", "Lawyers should supervise AI output with active review and verification analogous to supervising a junior associate", "Cites ABA Formal Opinion 512", "Cautions against free consumer AI tools in favour of enterprise-grade systems with security protocols", "Document not opened in this research run" ], "relatedIncidents": [], "sources": [ { "label": "LASC Technology Commission Generative AI Guidelines (PDF)", "url": "https://lajudicialcollege.org/LASC%20Tech%20Commission%20Generative%20AI%20Guidelines.pdf" } ], "country": "US", "categories": [ "judicial-use", "bar-guidance", "confidentiality-client-data", "verification-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "canada-federal-court-interim-ai-principles", "url": "https://safelegalai.com/regulation/canada/canada-federal-court-interim-ai-principles", "title": "Interim Principles and Guidelines on the Court's Use of Artificial Intelligence", "body": "Federal Court of Canada", "bodyShort": "Federal Court of Canada", "jurisdiction": "canada", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "disclosure" ], "effectiveDate": "2025-09-29T00:00:00.000Z", "versions": [ { "version": "Interim principles", "date": "2025-09-29T00:00:00.000Z", "url": "https://www.fct-cf.ca/en/pages/law-and-practice/artificial-intelligence" } ], "summary": "Interim principles published by the Federal Court of Canada on 29 September 2025 governing the court's own use of AI. The court commits not to use AI or automated decision-making in determining the outcome of cases without prior public consultation, and sets accountability, transparency and confidentiality requirements for administrative uses by judges and staff.", "keyProvisions": [ "The court will not use AI or automated decision-making to decide cases without public consultation.", "Administrative uses of AI by judges and staff must be accountable and transparent.", "Confidential and personal information in court records must be protected in any AI use." ], "relatedIncidents": [], "sources": [ { "label": "Federal Court of Canada — Artificial intelligence page", "url": "https://www.fct-cf.ca/en/pages/law-and-practice/artificial-intelligence" } ], "country": "CA", "categories": [ "court-own-use", "judicial-use", "ai-decision-prohibition", "confidentiality-client-data" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "it-legge-132-2025-intelligenza-artificiale", "url": "https://safelegalai.com/regulation/other/it-legge-132-2025-intelligenza-artificiale", "title": "Law 23 September 2025 no. 132 on artificial intelligence (Legge 132/2025)", "body": "Italian Parliament (Parlamento italiano)", "bodyShort": "Parliament (IT)", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "firms", "all" ], "requirements": [ "disclosure", "permissive", "prohibition", "supervision", "competence" ], "effectiveDate": "2025-10-10T00:00:00.000Z", "versions": [ { "version": "Original (GU Serie Generale n. 223, 25 September 2025)", "date": "2025-09-25T00:00:00.000Z", "url": "https://www.gazzettaufficiale.it/eli/id/2025/09/25/25G00143/sg", "changes": "Enacted; in force 10 October 2025." } ], "summary": "Italy's national AI framework law, in force 10 October 2025 and read in conformity with the EU AI Act. Article 13 confines AI in the intellectual professions to support tasks and requires clients to be told in clear language which AI systems a professional uses. Article 15 reserves every judicial decision to the magistrate.", "keyProvisions": [ "Art. 13(1): AI in intellectual professions only for instrumental and support activities, with prevalence of the professional's intellectual work.", "Art. 13(2): information on the AI systems used must be communicated to the client in clear, simple and exhaustive language.", "Art. 15(1): decisions on interpretation and application of law, evaluation of facts and evidence and adoption of measures are always reserved to the magistrate.", "Art. 15(2)-(3): Ministry of Justice regulates and, pending full AI Act application, authorises AI experimentation and use in ordinary judicial offices after consulting AgID and ACN.", "Art. 15(4): the Minister of Justice promotes AI training for magistrates and administrative staff.", "Art. 20: AgID and ACN designated as national AI authorities." ], "relatedIncidents": [], "sources": [ { "label": "Gazzetta Ufficiale — Legge 132/2025", "url": "https://www.gazzettaufficiale.it/eli/id/2025/09/25/25G00143/sg" }, { "label": "Normattiva consolidated text", "url": "https://www.normattiva.it/uri-res/N2Ls?urn:nir:stato:legge:2025-09-23;132" } ], "country": "IT", "categories": [ "client-consent-transparency", "judicial-use", "ai-decision-prohibition", "court-own-use", "ai-act-horizontal", "training-cle", "procurement-vendor", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "au-qld-sc-pd-5-2025-accuracy-of-references", "url": "https://safelegalai.com/regulation/australia/au-qld-sc-pd-5-2025-accuracy-of-references", "title": "Practice Direction Number 5 of 2025: Accuracy of References in Submissions", "body": "Supreme Court of Queensland", "bodyShort": "Supreme Court of Queensland", "jurisdiction": "australia", "region": "Queensland", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "barristers", "litigants", "parties" ], "requirements": [ "certification", "verification", "record-keeping" ], "effectiveDate": "2025-09-24T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-09-24T00:00:00.000Z", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/882064/sc-pd-5-pf-2025.pdf" } ], "summary": "Issued by Chief Justice Bowskill on 24 September 2025, this practice direction responds to generative AI by requiring every written submission to name the individual person responsible for its contents. Naming oneself certifies that all references have been verified and that the document reflects the practitioner's own professional judgment.", "keyProvisions": [ "Written submissions must name the responsible person or persons at the end of the document; naming a firm is not sufficient.", "The person making oral submissions is the responsible person for those submissions and also for the written submissions.", "The responsible person must verify the accuracy and relevance of all references to legislation, authorities and other sources.", "Practitioners responsible for submissions containing non-existent material may be referred to the Legal Services Commissioner or required to show cause against a personal costs order.", "Self-represented litigants must check references against AustLII, Queensland Judgments and the official legislation sites; reliance on fake material may cause an adjournment and adverse costs.", "Expressly cites Ayinde v Haringey [2025] EWHC 1383 at [5] and the Queensland Law Society's AI guidance statement." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Queensland — Practice Direction 5 of 2025 (official PDF)", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/882064/sc-pd-5-pf-2025.pdf" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record", "litigant-in-person", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-id-oah-alj-ai-guidelines", "url": "https://safelegalai.com/regulation/us-state/us-id-oah-alj-ai-guidelines", "title": "Idaho Office of Administrative Hearings: Guidelines for Administrative Law Judges Regarding the Use of Artificial Intelligence", "body": "Idaho Office of Administrative Hearings", "bodyShort": "Idaho OAH", "jurisdiction": "us-state", "region": "Idaho", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff", "parties" ], "requirements": [ "prohibition", "disclosure", "confidentiality", "verification" ], "versions": [ { "version": "Original", "date": "2025-09-24T00:00:00.000Z", "url": "https://oah.idaho.gov/wp-content/uploads/2025/09/OAH-AI-Guidelines.pdf", "changes": "Signed by Chief ALJ Bryan Nickels" } ], "summary": "Idaho's administrative judiciary adopted a near-total ban on adjudicative AI use. ALJs may not use AI to draft orders or decisions, to research law, or to summarise evidence and party submissions; limited transcript use is allowed; and ALJs may require parties to certify that AI-drafted content was checked by a human.", "keyProvisions": [ "ALJs are not permitted to use AI or generative AI to write orders or decisions, at any stage of drafting including outlines", "ALJs are not permitted to use AI or generative AI for legal research in case proceedings", "ALJs are not permitted to use AI or generative AI to analyse or summarise evidence or party submissions, partly because uploaded data including PII may be retained for training", "AI-generated voice-to-text transcripts from licensed non-retaining software are permitted but may not be cited; citations must be to the recording and verified", "ALJs may request AI disclosure statements from represented and self-represented parties at their discretion, with recommended certification language", "Governing canons: Idaho Code of Conduct for ALJs 1.2, 2.2 and 2.5" ], "relatedIncidents": [], "sources": [ { "label": "OAH AI Guidelines (PDF)", "url": "https://oah.idaho.gov/wp-content/uploads/2025/09/OAH-AI-Guidelines.pdf" } ], "country": "US", "categories": [ "judicial-use", "disclosure-filings", "confidentiality-client-data", "court-own-use", "evidence-admissibility" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "at-oerak-ki-leitfaden-anwaltskanzleien", "url": "https://safelegalai.com/regulation/other/at-oerak-ki-leitfaden-anwaltskanzleien", "title": "Künstliche Intelligenz (KI) in Anwaltskanzleien — Leitfaden und Checkliste für KI-Dienstleister", "body": "Österreichischer Rechtsanwaltskammertag (Austrian Bar), Ausschuss IT und Digitalisierung", "bodyShort": "ÖRAK", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms", "providers" ], "requirements": [ "confidentiality", "verification", "competence", "supervision", "record-keeping" ], "versions": [ { "version": "Stand September 2025", "date": "2025-09-23T00:00:00.000Z", "url": "https://www.anwaltsblatt.at/fileadmin/user_upload/Anwaltsblatt/PDF/oerak_25_leitfaden-KI-folder_a4_250923.pdf", "changes": "Issued by the ÖRAK's IT and digitalisation committee and distributed with the Anwaltsblatt." } ], "summary": "The Austrian Bar's four-page guide of September 2025 on using AI in law firms, with an annexed checklist that an AI provider must sign before a firm may instruct it. It applies § 9(2) RAO confidentiality and the binding § 40(3) RL-BA service-provider requirements to AI suppliers and makes staff AI competence a precondition of use.", "keyProvisions": [ "Entering mandate-related or confidential information into public or unsecured AI systems breaks confidentiality (§ 9(2) RAO) and is professionally impermissible.", "Lawyers may instruct only AI providers that meet § 40(3) RL-BA; otherwise only abstract, anonymous questions without any client link may be asked.", "The provider must confirm in writing: no use of firm input for model training, secure deletion, an Art. 28 GDPR processor agreement, EU or adequacy-country hosting (e.g. Switzerland, UK), equivalent compliance by sub-processors, and immediate notification if searched by police.", "The police-notification duty falls away for sub-processors that store data only briefly and process it fully automatically without human access (§ 40(3)(4) RL-BA).", "'Die anwaltliche Sorgfaltspflicht erfordert eine ausnahmslose Überprüfung und Verifizierung aller KI-generierten Ergebnisse'; final responsibility stays with the lawyer.", "AI use is permitted only where staff have AI competence, maintained by regular training, under a binding internal directive naming permitted tools, prohibited data and a responsible person." ], "relatedIncidents": [], "sources": [ { "label": "ÖRAK — Künstliche Intelligenz (KI) in Anwaltskanzleien (official PDF via Anwaltsblatt)", "url": "https://www.anwaltsblatt.at/fileadmin/user_upload/Anwaltsblatt/PDF/oerak_25_leitfaden-KI-folder_a4_250923.pdf" } ], "country": "AT", "categories": [ "bar-guidance", "procurement-vendor", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "training-cle", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pe-ley-31814-inteligencia-artificial", "url": "https://safelegalai.com/regulation/other/pe-ley-31814-inteligencia-artificial", "title": "Ley 31814: Law promoting the use of artificial intelligence for the country's economic and social development", "body": "Congress of the Republic of Peru (Congreso de la Republica)", "bodyShort": "Congreso", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification", "permissive" ], "effectiveDate": "2023-07-05T00:00:00.000Z", "versions": [ { "version": "Reglamento, Decreto Supremo 115-2025-PCM", "date": "2025-09-09T00:00:00.000Z", "url": "https://busquedas.elperuano.pe/dispositivo/NL/2192926-1", "changes": "Implementing regulation adding a risk-based approach and obligations; the exact URL of the decree was not verified." }, { "version": "Published in El Peruano", "date": "2023-07-05T00:00:00.000Z", "url": "https://busquedas.elperuano.pe/dispositivo/NL/2192926-1" } ], "summary": "Latin America's first AI statute, published 5 July 2023: it promotes AI within Peru's digital transformation under principles of human dignity and rights, names the PCM Secretariat of Government and Digital Transformation as national authority, declares justice among priority sectors, and requires annual reports to Congress. The 2025 implementing decree introduces risk classification.", "keyProvisions": [ "Art. 1: promotion of AI in a secure, ethical and responsible environment.", "Art. 2: national interest in AI for education, health, justice and security.", "Art. 4: PCM-SGTD as national technical-regulatory authority.", "Reglamento DS 115-2025-PCM adds risk-based obligations from September 2025." ], "relatedIncidents": [], "sources": [ { "label": "El Peruano — Ley 31814", "url": "https://busquedas.elperuano.pe/dispositivo/NL/2192926-1" } ], "country": "PE", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-ciarb-ai-arbitration-guideline", "url": "https://safelegalai.com/regulation/international/int-ciarb-ai-arbitration-guideline", "title": "Ciarb Guideline on the Use of AI in Arbitration (2025)", "body": "Chartered Institute of Arbitrators", "bodyShort": "Ciarb", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "disclosure", "verification", "confidentiality", "competence", "record-keeping" ], "versions": [ { "version": "Updated September 2025", "date": "2025-09-05T00:00:00.000Z", "url": "https://www.ciarb.org/media/bpndtcgu/guideline-on-the-use-of-ai-in-arbitration_updated-sept-2025.pdf", "changes": "Revised text; adds a template Agreement on the Use of AI and a template Procedural Order." }, { "version": "Original", "date": "2025-03-01T00:00:00.000Z", "url": "https://www.ciarb.org/news-listing/ciarb-launches-guideline-on-the-use-of-ai-in-arbitration/" } ], "summary": "Ciarb's non-binding guideline, launched in March 2025 and updated in September 2025, in four parts: benefits and risks; general recommendations; arbitrators' powers to rule on and require disclosure of AI use; and arbitrators' own use of AI. Appendices give a template party agreement and procedural order. It adopts the OECD definition of AI.", "keyProvisions": [ "Art. 7: tribunals may impose continuing AI-disclosure duties on parties, experts and witnesses; non-disclosure may lead to adverse inferences or costs consequences.", "Art. 6: tribunals may rule on admissibility of AI-generated material, consider bias and training data, and require the link between record and output to be explained.", "Art. 8.2: arbitrators should not relinquish decision-making to AI or delegate legal analysis and application of law to facts.", "Art. 8.3-8.4: arbitrators must independently verify AI-derived information and assume responsibility for the whole award.", "Art. 9: arbitrators should consult parties before using an AI tool and refrain if parties disagree." ], "relatedIncidents": [], "sources": [ { "label": "Ciarb — launch announcement", "url": "https://www.ciarb.org/news-listing/ciarb-launches-guideline-on-the-use-of-ai-in-arbitration/" }, { "label": "Ciarb — Guideline, updated September 2025 (PDF)", "url": "https://www.ciarb.org/media/bpndtcgu/guideline-on-the-use-of-ai-in-arbitration_updated-sept-2025.pdf" } ], "country": "INT", "categories": [ "arbitration-adr", "disclosure-filings", "verification-duty", "evidence-admissibility", "ai-decision-prohibition", "confidentiality-client-data", "sanctions-record", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "no-advokatforeningen-ai-guidance", "url": "https://safelegalai.com/regulation/other/no-advokatforeningen-ai-guidance", "title": "Veiledning for bruk av kunstig intelligens i advokatvirksomhet", "body": "Norwegian Bar Association [Advokatforeningen]", "bodyShort": "Advokatforeningen", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "permissive" ], "versions": [ { "version": "Approved by the board", "date": "2025-09-05T00:00:00.000Z", "url": "https://www.advokatforeningen.no/advokatetikk/anbefalinger/veiledning-for-bruk-av-kunstig-intelligens-i-advokatvirksomhet/", "changes": "Prepared by the Association's professional ethics committee and approved by the board on 5 September 2025; full text is members-only. Full text is members-only and was not read, so this record is unverified." } ], "summary": "Guidance produced by the Norwegian Bar Association's professional ethics committee and approved by its board on 5 September 2025, on using AI in legal practice. It is issued under the Association's ethics recommendations and holds that the lawyer always carries final responsibility even where AI is used as a tool.", "keyProvisions": [ "Issued under Advokatforeningen's professional-ethics recommendations for lawyers and law firms.", "The lawyer retains final responsibility for AI-assisted work.", "Full text is behind a member login; substantive provisions could not be verified from public sources." ], "relatedIncidents": [], "sources": [ { "label": "Advokatforeningen — Veiledning for bruk av kunstig intelligens i advokatvirksomhet", "url": "https://www.advokatforeningen.no/advokatetikk/anbefalinger/veiledning-for-bruk-av-kunstig-intelligens-i-advokatvirksomhet/" } ], "country": "NO", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ca-qc-superior-court-ai-governance-framework", "url": "https://safelegalai.com/regulation/canada/ca-qc-superior-court-ai-governance-framework", "title": "Cadre de gouvernance en matière d'intelligence artificielle [Artificial Intelligence Governance Framework]", "body": "Superior Court of Quebec [Cour supérieure du Québec]", "bodyShort": "Quebec Superior Court", "jurisdiction": "canada", "region": "Quebec", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "supervision", "competence" ], "versions": [ { "version": "September 2025", "date": "2025-09-01T00:00:00.000Z", "url": "https://coursuperieureduquebec.ca/fileadmin/cour-superieure/A_propos/2025-09_Cadre_gouvernance_IA_Cour_superieure_Quebec.pdf", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the court's published framework listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A governance framework adopted by the Superior Court of Quebec in September 2025 for the court's own AI use, alongside a chatbot pilot as a judicial support tool. It sets six principles: transparency and public confidence, judicial independence, support for judicial functions, ethics and professional conduct, caution, and responsibility, and requires rigorous verification of AI-generated content.", "keyProvisions": [ "AI supports judicial functions; it does not replace judicial deliberation or decision.", "Six governing principles: transparency and public confidence, judicial independence, support for judicial functions, ethics, caution, responsibility.", "Content generated by AI must always be rigorously verified by the judge or support staff using it, because of hallucination risk." ], "relatedIncidents": [], "sources": [ { "label": "Cour supérieure du Québec — Cadre de gouvernance IA (PDF)", "url": "https://coursuperieureduquebec.ca/fileadmin/cour-superieure/A_propos/2025-09_Cadre_gouvernance_IA_Cour_superieure_Quebec.pdf" }, { "label": "FJA — repository of AI initiatives in courts and tribunals", "url": "https://www.fja.gc.ca/COVID-19/Repository-of-AI-Initiatives-in-Courts-and-Tribunals-Repertoire-dinitiatives-IA-dans-les-tribunaux-eng.html" } ], "country": "CA", "categories": [ "court-own-use", "judicial-use", "ai-decision-prohibition", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ca-rule-of-court-10-430-generative-ai", "url": "https://safelegalai.com/regulation/us-state/ca-rule-of-court-10-430-generative-ai", "title": "California Rule of Court 10.430 and Standard 10.80: Generative AI use by courts", "body": "Judicial Council of California", "bodyShort": "Judicial Council of California", "jurisdiction": "us-state", "region": "California", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "confidentiality", "verification", "disclosure", "record-keeping" ], "effectiveDate": "2025-09-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-09-01T00:00:00.000Z", "url": "https://courts.ca.gov/cms/rules/index/ten/rule10_430" } ], "summary": "Rule 10.430 of the California Rules of Court, effective 1 September 2025, requires every California court that permits generative AI use by judicial officers or staff to adopt a use policy by 15 December 2025. Policies must address confidentiality, bias, human review of output, and disclosure when public-facing content is fully AI-generated. Standard 10.80 offers a model policy.", "keyProvisions": [ "Courts permitting generative AI must adopt a written use policy by 15 December 2025.", "Policies must prohibit entering confidential, personal or sealed information into public tools.", "Staff must review AI output for accuracy and bias before use.", "Public-facing written or visual content generated entirely by AI must be disclosed as such." ], "relatedIncidents": [], "sources": [ { "label": "California Courts — Rule 10.430", "url": "https://courts.ca.gov/cms/rules/index/ten/rule10_430" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "jp-jfba-generative-ai-notice-2025", "url": "https://safelegalai.com/regulation/other/jp-jfba-generative-ai-notice-2025", "title": "Points of Attention on the Use of Generative AI in Attorney Practice: Five Points for Appropriate Use", "body": "Japan Federation of Bar Associations (Nichibenren), AI Strategy Working Group", "bodyShort": "JFBA", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "confidentiality", "verification", "competence" ], "effectiveDate": "2025-09-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-09-01T00:00:00.000Z", "url": "https://olga-legal.com/13865/", "changes": "Published September 2025 on the JFBA members-only site; exact day not public and the text is not publicly accessible." } ], "summary": "Members-only guidance issued in September 2025 by the JFBA's AI Strategy Working Group setting five points for appropriate use of generative AI in legal practice: keep confidential and client-identifiable information out of external AI services; treat outputs including cited cases as unverified until checked against primary sources; the attorney bears final responsibility.", "keyProvisions": [ "Confidentiality: avoid inputting client-identifiable or confidential information into external generative-AI services.", "Hallucination: verify AI-cited statutes and case law against primary sources before use.", "Responsibility remains with the attorney; AI output is not exculpatory.", "Client explanation and record-keeping of verification steps." ], "relatedIncidents": [], "sources": [ { "label": "Seminar page quoting the title, working group, date and members-only status", "url": "https://olga-legal.com/13865/" }, { "label": "Practitioner commentary on the JFBA notice", "url": "https://hidetoshilawoffice.com/column/column-20260627-jfba-ai-guidelines/" } ], "country": "JP", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "mx-collegiate-court-ai-judicial-criteria-2025", "url": "https://safelegalai.com/regulation/other/mx-collegiate-court-ai-judicial-criteria-2025", "title": "Federal collegiate court criteria on the use of artificial intelligence in judicial decision-making", "body": "Second Collegiate Court in Civil Matters of the Second Circuit [Segundo Tribunal Colegiado en Materia Civil del Segundo Circuito]", "bodyShort": "Mexican federal collegiate court", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "verification", "prohibition" ], "versions": [ { "version": "Published in the Semanario Judicial de la Federación", "date": "2025-08-22T00:00:00.000Z", "url": "https://elmundodelderecho.com/scjn-fija-lineamientos-para-uso-etico-de-inteligencia-artificial-en-procesos-judiciales/", "changes": "Source is legal press only; the criteria were not verified against the official Semanario Judicial de la Federacion text, so this record is unverified." } ], "summary": "Criteria published in Mexico's Semanario Judicial de la Federacion on 22 August 2025 holding that AI may assist a judge only as an auxiliary tool, under an exercise of self-restraint and with respect for human rights, and that deliberation and the final decision remain strictly within the jurisdictional sphere.", "keyProvisions": [ "Technology operates as an auxiliary and not as a substitute; deliberation and decision remain in the strictly jurisdictional sphere.", "The final decision in a judicial process must rest exclusively with judges and magistrates.", "Reported guiding principles: proportionality and non-harm, protection of personal data, transparency and explainability, and human judicial supervision." ], "relatedIncidents": [], "sources": [ { "label": "El Mundo del Derecho — SCJN criteria on ethical AI use in judicial proceedings", "url": "https://elmundodelderecho.com/scjn-fija-lineamientos-para-uso-etico-de-inteligencia-artificial-en-procesos-judiciales/" } ], "country": "MX", "categories": [ "judicial-use", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ke-judiciary-draft-ai-policy", "url": "https://safelegalai.com/regulation/other/ke-judiciary-draft-ai-policy", "title": "Draft Judiciary Artificial Intelligence Policy and draft Practice Directions on AI use in courts", "body": "Judiciary of Kenya", "bodyShort": "Judiciary of Kenya", "jurisdiction": "other", "type": "policy", "status": "consultation", "appliesTo": [ "judges", "court-staff", "lawyers", "litigants" ], "requirements": [ "certification", "verification", "disclosure", "supervision", "risk-classification" ], "versions": [ { "version": "Announcement of the AI Adoption Policy Framework", "date": "2025-08-11T00:00:00.000Z", "url": "https://judiciary.go.ke/judiciary-to-leverage-ai-to-enhance-justice/", "changes": "Chief Justice Martha Koome confirms the Judiciary is developing an AI Adoption Policy Framework covering case management, legal research, predictive analytics and administrative support. The draft policy text is not published; the record rests on the Judiciary's own announcement and press, so it is unverified." } ], "summary": "Kenya's Judiciary is developing an AI Adoption Policy Framework, announced officially in August 2025, and circulated a draft AI Policy with draft Practice Directions for public participation in May 2026. Reporting describes a three-tier risk model, a certificate of human verification for AI-assisted filings, and a bar on AI displacing judicial decision-making. Not in force.", "keyProvisions": [ "Announced aim: guide integration of AI tools while safeguarding judicial independence, data privacy and due process (Judiciary of Kenya, 11 Aug 2025).", "Reported draft: any AI-assisted filing must be accompanied by a certificate of human verification, placing responsibility on the advocate rather than the tool.", "Reported draft: three risk tiers — bail assessment and predictive sentencing require mandatory human oversight and audits; legal research requires verification and disclosure; scheduling and transcription face general monitoring.", "Reported draft: a clear boundary between administrative automation and adjudicative discretion, with AI as decision support only." ], "relatedIncidents": [], "sources": [ { "label": "Judiciary to leverage AI to enhance justice (Judiciary of Kenya official, 11 Aug 2025)", "url": "https://judiciary.go.ke/judiciary-to-leverage-ai-to-enhance-justice/" }, { "label": "Judiciary to introduce AI policy amid concerns of bot-generated filings (Daily Nation)", "url": "https://nation.africa/kenya/news/judiciary-to-introduce-ai-policy-amid-concerns-of-bot-generated-filings--5475558" }, { "label": "Judiciary moves to regulate AI use in courtrooms (Techweez, 15 May 2026)", "url": "https://techweez.com/2026/05/15/kenya-judiciary-ai-policy/" } ], "country": "KE", "categories": [ "disclosure-filings", "verification-duty", "judicial-use", "ai-decision-prohibition", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-eoir-pm-25-40-generative-ai", "url": "https://safelegalai.com/regulation/us-federal/us-eoir-pm-25-40-generative-ai", "title": "Policy Memorandum 25-40: Use of Generative Artificial Intelligence in EOIR Proceedings", "body": "United States Department of Justice, Executive Office for Immigration Review", "bodyShort": "DOJ EOIR", "jurisdiction": "us-federal", "type": "policy", "status": "in-force", "appliesTo": [ "lawyers", "judges", "litigants", "court-staff" ], "requirements": [ "permissive", "verification", "competence" ], "versions": [ { "version": "Original", "date": "2025-08-08T00:00:00.000Z", "url": "https://www.justice.gov/eoir/media/1410621/dl?inline=" } ], "summary": "Signed by Acting Director Sirce E. Owen and effective 8 August 2025, this is the first Department of Justice component policy on generative AI binding practitioners. EOIR imposes neither a blanket prohibition nor mandatory disclosure, but treats hallucinated or erroneous AI content as a likely professional-conduct violation and directs adjudicators to refer suspected misuse for discipline.", "keyProvisions": [ "EOIR has no blanket prohibition on generative AI and no mandatory disclosure requirement in its proceedings.", "Individual immigration judges, courts and the Board of Immigration Appeals may adopt standing orders or local operating procedures on AI use and disclosure.", "Practitioners filing hallucinated or erroneous AI-generated content likely violate professional conduct obligations and may be disciplined under 8 C.F.R. 1003.102(c) for knowingly or recklessly offering false evidence.", "Adjudicators who suspect improper AI use should refer the matter to EOIR's Attorney Discipline Program and, as appropriate, its Anti-Fraud Program.", "Improper use of generative AI by an EOIR adjudicator may result in corrective or disciplinary action; unauthorised AI services on DOJ government-furnished equipment are prohibited.", "Cites OCAHO's guidance established by adjudication in United States v. Wallcon, LLC, 21 OCAHO no. 1630 (2025), and OMB Memorandum M-25-21." ], "relatedIncidents": [], "sources": [ { "label": "DOJ EOIR — PM 25-40 (official PDF)", "url": "https://www.justice.gov/eoir/media/1410621/dl?inline=" } ], "country": "US", "categories": [ "bar-guidance", "disclosure-filings", "sanctions-record", "verification-duty", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ar-rosario-camara-giacomino-2025", "url": "https://safelegalai.com/regulation/other/ar-rosario-camara-giacomino-2025", "title": "Camara de Apelacion en lo Civil y Comercial de Rosario, Sala II, Giacomino c/ Monserrat: AI-fabricated citations", "body": "Civil and Commercial Court of Appeals of Rosario (Camara de Apelacion en lo Civil y Comercial de Rosario)", "bodyShort": "Camara Rosario", "jurisdiction": "other", "region": "Santa Fe", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "verification" ], "effectiveDate": "2025-08-01T00:00:00.000Z", "versions": [ { "version": "Judgment (CUIJ 21-11893083-2, August 2025)", "date": "2025-08-01T00:00:00.000Z", "url": "https://abogados.com.ar/la-primera-sentencia-argentina-contra-la-ia-el-riesgo-de-una-profesion-en-piloto-automatico/37412", "changes": "Initial decision; reported by Abogados.com.ar, official text not obtained." } ], "summary": "Argentina's first appellate pronouncement on AI hallucinations: after the lawyer admitted the cited cases were invented by a chatbot, the court held that delegating research to AI and filing unverified citations is extremely risky and even reckless, declined a fine but referred the lawyer to the Rosario bar for disciplinary review and preventive measures.", "keyProvisions": [ "No consent relieves a lawyer of the duty to verify sources.", "Filing AI-fabricated jurisprudence breaches professional ethics regardless of good faith.", "Referral to the Colegio de Abogados de Rosario.", "Followed by General Roca (26 fabricated citations) and Neuquen referrals." ], "relatedIncidents": [], "sources": [ { "label": "Abogados.com.ar analysis", "url": "https://abogados.com.ar/la-primera-sentencia-argentina-contra-la-ia-el-riesgo-de-una-profesion-en-piloto-automatico/37412" }, { "label": "Premisa report", "url": "https://www.premisa.com.ar/article/primer_sentencia_argentina_que_marca_un_limite_etico_al_uso_de_ia_en_la_abogacia2025725" } ], "country": "AR", "categories": [ "sanctions-record", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "fr-ministere-justice-rapport-ia-justice", "url": "https://safelegalai.com/regulation/other/fr-ministere-justice-rapport-ia-justice", "title": "L'IA au service de la justice : stratégie et solutions opérationnelles", "body": "Ministère de la Justice (Ministry of Justice), France", "bodyShort": "Ministère de la Justice", "jurisdiction": "other", "type": "report", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "competence" ], "versions": [ { "version": "Working group report", "date": "2025-08-01T00:00:00.000Z", "url": "https://www.justice.gouv.fr/sites/default/files/2025-08/rapport_ia_au_service_de_la_justice.pdf", "changes": "Report of the Ministry of Justice working group on AI in the service of justice. Month-only date: the source gives the month, so the 1st is recorded as the date. The full report was not read; the record rests on the ministry's published PDF listing and announcement, so it is unverified." } ], "summary": "The French Ministry of Justice's working-group report, published August 2025, on deploying AI across the judicial order. It follows the February 2025 announcement of four priority use cases and sets out safeguards — a charter of AI use for staff, systematic verification of AI output, attention to bias, protection of confidential data and preference for nationally hosted solutions.", "keyProvisions": [ "Four priority use cases: transcription of interviews, legal research and access to case law, interpretation and translation, and summarising large document volumes.", "A charter of AI use for Ministry staff requiring 'vérification et contrôle systématique des résultats fournis par l'IA'.", "Preference for solutions hosted on national territory, with legal frameworks for data protection compliance.", "A sovereign internal generative-AI assistant for magistrates and Ministry staff, with 'Mon Assistant Pénal' and 'Mon Assistant Civil' developed with the Paris Court of Appeal and DINUM.", "Training modules on AI and data handling available to all staff." ], "relatedIncidents": [], "sources": [ { "label": "Ministère de la Justice — rapport IA au service de la justice (official PDF)", "url": "https://www.justice.gouv.fr/sites/default/files/2025-08/rapport_ia_au_service_de_la_justice.pdf" }, { "label": "Ministère de la Justice — Développer l'IA pour la Justice (official)", "url": "https://www.justice.gouv.fr/actualites/actualite/developper-lintelligence-artificielle-ia-justice-entre-innovation-securisation" } ], "country": "FR", "categories": [ "court-own-use", "judicial-use", "training-cle", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-la-act-250-2025-ai-evidence", "url": "https://safelegalai.com/regulation/us-state/us-la-act-250-2025-ai-evidence", "title": "Louisiana Act No. 250 of 2025: verification and disclosure of AI-generated or falsified evidence (La. C.C.P. art. 371)", "body": "Louisiana Legislature", "bodyShort": "Louisiana Legislature", "jurisdiction": "us-state", "region": "Louisiana", "type": "statute", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification", "disclosure", "certification" ], "effectiveDate": "2025-08-01T00:00:00.000Z", "versions": [ { "version": "Act No. 250 (2025 Regular Session, HB 178)", "date": "2025-08-01T00:00:00.000Z", "url": "https://www.legis.la.gov/legis/ViewDocument.aspx?d=1425558", "changes": "Signed 11 June 2025, effective 1 August 2025" } ], "summary": "Louisiana enacted a civil-procedure duty to police AI-falsified evidence. Counsel must exercise reasonable diligence to verify exhibit authenticity, disclose known falsification of their own exhibits including AI generation or alteration, and raise reasonable suspicion about an opponent's exhibits at the pretrial stage, on pain of contempt and discipline.", "keyProvisions": [ "Attorneys must exercise reasonable diligence to verify the authenticity of evidence before offering it to the court", "A party that knows or has reason to know its own exhibits have been falsified, including by AI generation or alteration, must disclose that fact", "A party with reasonable suspicion that opposing exhibits are AI-generated or falsified must raise it at the pretrial conference or a pretrial admissibility hearing", "Counsel offering false or AI-manipulated evidence without disclosure face contempt and potential disciplinary action where they knew or should have known through reasonable diligence", "Applies to civil proceedings only; does not address AI-drafted briefs containing hallucinated citations", "Enrolled text not opened in this research run and must be confirmed before publication" ], "relatedIncidents": [], "sources": [ { "label": "Act 250 enrolled text (official)", "url": "https://www.legis.la.gov/legis/ViewDocument.aspx?d=1425558" } ], "country": "US", "categories": [ "evidence-admissibility", "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-ao-judiciary-interim-ai-guidance", "url": "https://safelegalai.com/regulation/us-federal/us-ao-judiciary-interim-ai-guidance", "title": "Interim guidance on the use of artificial intelligence in the federal judiciary", "body": "Administrative Office of the United States Courts, Artificial Intelligence Task Force", "bodyShort": "AO of the US Courts", "jurisdiction": "us-federal", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "disclosure", "confidentiality", "supervision" ], "versions": [ { "version": "Interim guidance, distributed judiciary-wide", "date": "2025-07-31T00:00:00.000Z", "url": "https://www.uscourts.gov/data-news/reports/annual-reports/directors-annual-report/annual-report-2025/court-operations-annual-report-2025", "changes": "First judiciary-wide AI guidance; temporary pending a permanent AI governance framework called for by the September 2025 Strategic Plan." } ], "summary": "Interim guidance developed by the Administrative Office's AI Task Force and distributed across the United States federal judiciary on 31 July 2025. It permits experimentation with AI while directing users to independently verify all AI-generated output, cautioning against delegating core judicial functions, and asking each court to define locally approved tools and tasks. The text itself is not published.", "keyProvisions": [ "Judiciary users must review and independently verify all AI-generated content and remain accountable for AI-assisted work.", "Cautions against delegating core judicial functions to AI, including decision making and case adjudication.", "Advises extreme caution when using AI on novel legal questions.", "Courts should define the tasks for which locally approved AI tools may be used.", "Users should consider whether AI use should be disclosed, to whom, and consistently with confidentiality and ethical duties." ], "relatedIncidents": [], "sources": [ { "label": "US Courts — Court Operations, Director's Annual Report 2025", "url": "https://www.uscourts.gov/data-news/reports/annual-reports/directors-annual-report/annual-report-2025/court-operations-annual-report-2025" }, { "label": "Administrative Office letter to Senator Grassley on judiciary use of AI (PDF)", "url": "https://www.grassley.senate.gov/imo/media/doc/ao_to_grassley_re_judiciary_use_of_ai.pdf" } ], "country": "US", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "kerala-hc-ai-policy-district-judiciary", "url": "https://safelegalai.com/regulation/india/kerala-hc-ai-policy-district-judiciary", "title": "Policy Regarding Use of Artificial Intelligence Tools in District Judiciary", "body": "High Court of Kerala", "bodyShort": "Kerala High Court", "jurisdiction": "india", "region": "Kerala", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "confidentiality" ], "effectiveDate": "2025-07-19T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-07-19T00:00:00.000Z", "url": "https://images.assettype.com/theleaflet/2025-07-22/mt4bw6n7/Kerala_HC_AI_Guidelines.pdf", "archiveUrl": "https://web.archive.org/web/20260904133706/https://images.assettype.com/theleaflet/2025-07-22/mt4bw6n7/Kerala_HC_AI_Guidelines.pdf" } ], "summary": "A policy issued by the High Court of Kerala on 19 July 2025 for the state's district judiciary, the first by an Indian High Court. Judges and court staff may use AI only for assistance and must not use it to reach findings, orders or judgments; outputs must be verified and confidential case data kept out of public tools.", "keyProvisions": [ "AI tools may not be used to arrive at findings, reliefs, orders or judgments.", "Approved AI may assist with administrative and clerical tasks only.", "Any AI output used must be verified by the judicial officer.", "Confidential case information must not be entered into public AI tools." ], "relatedIncidents": [ "pooja-ramesh-singh-v-jk-bank" ], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "Kerala High Court AI policy (PDF mirror hosted by The Leaflet)", "url": "https://images.assettype.com/theleaflet/2025-07-22/mt4bw6n7/Kerala_HC_AI_Guidelines.pdf" }, { "label": "High Court of Kerala", "url": "https://highcourt.kerala.gov.in/" } ], "country": "IN", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "eu-commission-guidelines-gpai-providers", "url": "https://safelegalai.com/regulation/eu/eu-commission-guidelines-gpai-providers", "title": "Guidelines on the scope of obligations for providers of general-purpose AI models", "body": "European Commission", "bodyShort": "European Commission", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers" ], "requirements": [ "record-keeping", "disclosure" ], "effectiveDate": "2025-07-18T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-07-18T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/policies/guidelines-gpai-providers" } ], "summary": "European Commission guidelines published 18 July 2025 clarifying which organisations are providers of general-purpose AI models under the EU AI Act and what documentation, copyright and transparency duties attach from 2 August 2025. The guidelines are interpretive and relevant to legal tech vendors that fine-tune or distribute foundation models in the EU.", "keyProvisions": [ "Defines when a model counts as general-purpose and when a downstream modifier becomes a provider.", "Explains documentation and information duties owed to downstream deployers.", "Sets out the Commission's approach to enforcement in the first year after 2 August 2025." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — GPAI provider guidelines", "url": "https://digital-strategy.ec.europa.eu/en/policies/guidelines-gpai-providers" } ], "country": "EU", "categories": [ "ai-act-horizontal", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "br-cam-ccbc-orientacao-administrativa-07-2025", "url": "https://safelegalai.com/regulation/brazil/br-cam-ccbc-orientacao-administrativa-07-2025", "title": "Orientacao Administrativa no 07/2025: Use of Artificial Intelligence Tools in Arbitrations", "body": "Brazil-Canada Arbitration and Mediation Centre (Centro de Arbitragem e Mediacao Brasil-Canada)", "bodyShort": "CAM-CCBC", "jurisdiction": "brazil", "type": "guidance", "status": "in-force", "appliesTo": [ "parties", "lawyers" ], "requirements": [ "confidentiality", "verification", "disclosure" ], "effectiveDate": "2025-07-17T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-07-17T00:00:00.000Z", "url": "https://www.ccbc.org.br/cam-ccbc-centro-arbitragem-mediacao/orientacao-administrativa-no-07-2025/" } ], "summary": "Guidance for arbitrations administered by CAM-CCBC: all participants must understand AI tools and their limits; AI use must respect confidentiality; party representatives are responsible for the factual and legal accuracy of AI-assisted work; disclosure is not required as a general rule but should be discussed case by case; and arbitrators may not delegate decision-making to AI.", "keyProvisions": [ "Participants are responsible for understanding AI tools and their limitations.", "Confidentiality and data-protection duties apply to AI use in arbitration.", "Representatives must review AI output for accuracy; no use of AI to disrupt proceedings or falsify evidence.", "Arbitrators retain their personal mandate; decision-making functions cannot be delegated.", "Disclosure decided case by case; CAM-CCBC itself does not use AI in administration." ], "relatedIncidents": [], "sources": [ { "label": "CAM-CCBC — Orientacao Administrativa 07/2025", "url": "https://www.ccbc.org.br/cam-ccbc-centro-arbitragem-mediacao/orientacao-administrativa-no-07-2025/" } ], "country": "BR", "categories": [ "arbitration-adr" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ar-cpacf-guia-ia-abogados-2025", "url": "https://safelegalai.com/regulation/other/ar-cpacf-guia-ia-abogados-2025", "title": "Guide for the use of Artificial Intelligence for Lawyers (Guia para el uso de Inteligencia Artificial para Abogados)", "body": "Public Bar of the Federal Capital (Colegio Publico de la Abogacia de la Capital Federal)", "bodyShort": "CPACF", "jurisdiction": "other", "region": "Ciudad Autonoma de Buenos Aires", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "consent" ], "effectiveDate": "2025-07-11T00:00:00.000Z", "versions": [ { "version": "Original (Laboratorio Fintech y Legaltech, July 2025; published 6 August 2025)", "date": "2025-07-11T00:00:00.000Z", "url": "https://aldiaargentina.microjuris.com/wp-content/uploads/2025/08/CPACF-IA-GUIA.pdf" } ], "summary": "Buenos Aires bar guide built on four pillars: responsible use minimising hallucinations and bias; permanent human control with the lawyer as final decision-maker; protection of personal data and professional secrecy, including client authorisation before sharing confidential information; and continuous training. It states that AI does not exempt the lawyer from responsibility for anything signed or filed.", "keyProvisions": [ "Verify all AI outputs against referenced sources; beware hallucinations and false citations.", "No client data on platforms without security guarantees; obtain client authorisation where appropriate.", "Be transparent with clients about AI use and do not create false expectations.", "Do not delegate decisions to AI; the lawyer bears sole responsibility for filings.", "Continuous training, including prompt design." ], "relatedIncidents": [], "sources": [ { "label": "CPACF guide (PDF via Microjuris)", "url": "https://aldiaargentina.microjuris.com/wp-content/uploads/2025/08/CPACF-IA-GUIA.pdf" }, { "label": "Launch report (Microjuris)", "url": "https://aldiaargentina.microjuris.com/2025/08/06/actualidad-el-cpacf-presento-una-guia-para-el-uso-de-inteligencia-artificial-para-abogados/" } ], "country": "AR", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "client-consent-transparency", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eu-gpai-code-of-practice", "url": "https://safelegalai.com/regulation/eu/eu-gpai-code-of-practice", "title": "General-Purpose AI Code of Practice", "body": "European AI Office (independent chairs)", "bodyShort": "EU AI Office", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers" ], "requirements": [ "record-keeping", "disclosure", "permissive" ], "effectiveDate": "2025-07-10T00:00:00.000Z", "versions": [ { "version": "Final", "date": "2025-07-10T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai" } ], "summary": "A voluntary code drafted under the EU AI Office and published 10 July 2025 that gives providers of general-purpose AI models a route to demonstrate compliance with the AI Act's GPAI obligations. Its chapters cover transparency, copyright and safety and security. Commission enforcement powers over GPAI providers became effective on 2 August 2026.", "keyProvisions": [ "Transparency chapter sets a model documentation form for downstream providers and regulators.", "Copyright chapter commits signatories to respect rights reservations and lawful access to training data.", "Safety and security chapter applies to models with systemic risk.", "Signing is voluntary; signatories benefit from a presumption of good faith in enforcement." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — GPAI Code of Practice", "url": "https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai" } ], "country": "EU", "categories": [ "ai-act-horizontal", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "my-bar-circular-242-2025-generative-ai", "url": "https://safelegalai.com/regulation/other/my-bar-circular-242-2025-generative-ai", "title": "Circular No. 242/2025: Updates on the Use of Generative Artificial Intelligence in Legal Practice", "body": "Malaysian Bar [Badan Peguam Malaysia]", "bodyShort": "Malaysian Bar", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "record-keeping" ], "effectiveDate": "2025-07-03T00:00:00.000Z", "versions": [ { "version": "Original, with annexed advisory on the use of generative AI in legal practice", "date": "2025-07-03T00:00:00.000Z", "url": "https://www.malaysianbar.org.my/cms/upload_files/document/Circular%20No%20242-2025.pdf" } ], "summary": "The Malaysian Bar's principal AI instrument, issued 3 July 2025 by the Bar Council Cyber and Privacy Laws Committee and the Legal Tech, AI and Sandbox Committee. Its annexed advisory sets out recommended and prohibited uses, a nine-risk mitigation table and a four-R framework for practitioners adopting generative AI.", "keyProvisions": [ "All generative AI output must be independently verified against traditional legal databases; if verification is not possible the content must not be relied on.", "Do not input personal data, NRIC or financial data or client-specific facts; redact before use and comply with the Personal Data Protection Act 2010.", "All output must be subject to human oversight, legal judgment and ethical consideration.", "Keep clear records of professional oversight and final review for audit purposes.", "Read terms of use to establish whether inputs are retained or used for training; prefer legal-sector tools with privacy and regulatory controls.", "Identifies unauthorised practice of law as a generative AI risk; AI must not substitute for trained legal minds." ], "relatedIncidents": [], "sources": [ { "label": "Malaysian Bar — Circular No. 242/2025 (official PDF)", "url": "https://www.malaysianbar.org.my/cms/upload_files/document/Circular%20No%20242-2025.pdf" } ], "country": "MY", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "procurement-vendor", "training-cle", "unauthorised-practice" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ga-judicial-council-ai-committee-report", "url": "https://safelegalai.com/regulation/us-state/us-ga-judicial-council-ai-committee-report", "title": "Judicial Council of Georgia Ad Hoc Committee on Artificial Intelligence and the Courts: Report", "body": "Judicial Council of Georgia, Ad Hoc Committee on Artificial Intelligence and the Courts", "bodyShort": "Judicial Council of Georgia", "jurisdiction": "us-state", "region": "Georgia", "type": "report", "status": "in-force", "appliesTo": [ "judges", "court-staff", "lawyers" ], "requirements": [ "risk-classification", "verification" ], "versions": [ { "version": "Committee report", "date": "2025-07-03T00:00:00.000Z", "url": "https://www.gasupreme.us/07-03-2025-judicial-council-ai-committee-submits-report/", "changes": "Report submitted to the Judicial Council" } ], "summary": "Georgia's judicial branch response to generative AI. The Ad Hoc Committee on AI and the Courts, established August 2024 under Justice Andrew Pinson, submitted its report on 3 July 2025 focused on court rules and procedures; no implementing Supreme Court order has been verified as of September 2026.", "keyProvisions": [ "Committee established August 2024, chaired by Justice Andrew A. Pinson, with membership from state court classes and judicial partner agencies", "Charge: assess risks and benefits of generative AI in the courts and recommend measures to maintain public trust and confidence", "Principal focus on court rules and procedures", "A parallel State Bar of Georgia committee reviews the Georgia Rules of Professional Conduct", "Report submitted 3 July 2025; implementing order not verified" ], "relatedIncidents": [], "sources": [ { "label": "Judicial Council AI Committee submits report (official announcement)", "url": "https://www.gasupreme.us/07-03-2025-judicial-council-ai-committee-submits-report/" }, { "label": "Committee page", "url": "https://jcaoc.georgiacourts.gov/entity/ad-hoc-committee-on-artificial-intelligence-and-the-courts/" } ], "country": "US", "categories": [ "judicial-use", "disclosure-filings", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ph-senate-bill-25-ai-regulation", "url": "https://safelegalai.com/regulation/other/ph-senate-bill-25-ai-regulation", "title": "Senate Bill No. 25 (20th Congress): An Act Regulating the Development and Use of Artificial Intelligence Systems", "body": "Senate of the Philippines (Senator Pia S. Cayetano)", "bodyShort": "Senate of the Philippines", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "effectiveDate": "2025-07-02T00:00:00.000Z", "versions": [ { "version": "Filed", "date": "2025-07-02T00:00:00.000Z", "url": "https://senate.gov.ph/legacy/lis_bills/4649942502!.pdf" } ], "summary": "A pending Philippine Senate bill, filed 2 July 2025, proposing risk-based regulation of artificial-intelligence systems modelled on the European Union AI Act. It had not been enacted as at 4 September 2026 and does not separately classify judicial or legal-services uses of AI as high risk.", "keyProvisions": [ "Proposes a risk-based classification of AI systems modelled on the EU AI Act.", "Not enacted as at 4 September 2026; regulation of the Bench and Bar remains constitutionally reserved to the Supreme Court." ], "relatedIncidents": [], "sources": [ { "label": "Senate of the Philippines — Senate Bill No. 25 (official PDF)", "url": "https://senate.gov.ph/legacy/lis_bills/4649942502!.pdf" } ], "country": "PH", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "se-advokatsamfundet-generative-ai-guidance", "url": "https://safelegalai.com/regulation/other/se-advokatsamfundet-generative-ai-guidance", "title": "Vägledning om användning av generativ AI i advokatverksamhet", "body": "Swedish Bar Association [Sveriges advokatsamfund]", "bodyShort": "Swedish Bar Association", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "consent", "permissive" ], "versions": [ { "version": "Guidance to the board of the Swedish Bar Association", "date": "2025-06-12T00:00:00.000Z", "url": "https://www.advokatsamfundet.se/globalassets/advokatsamfundet_sv/advokatyrket/vagledning-om-anvandning-av-generativ-ai-i-advokatverksamhet.pdf", "changes": "Fourteen-page guidance produced by the working group that drafted the 2024 general advice." }, { "version": "Allmänna råd rörande användning av generativa AI-modeller i advokatverksamhet", "date": "2024-06-13T00:00:00.000Z", "url": "https://www.advokatsamfundet.se/globalassets/advokatsamfundet_sv/advokatyrket/allmanna-rad-rorande-anvandning-av-generativa-ai-modeller-i-advokatverksamhet.pdf", "changes": "General advice adopted by the board on 13 June 2024." } ], "summary": "Guidance dated Stockholm, 12 June 2025, from a Swedish Bar Association working group on using generative AI in legal practice. It ties GenAI use to the Code of Conduct (VRGA) across four areas: regulatory compliance, advisory liability, professional secrecy, and principal liability and office organisation.", "keyProvisions": [ "GenAI can produce fluent text containing factual errors or references to cases and statutes that do not exist; output must be critically reviewed before it enters advice.", "Advisory liability applies in full whether or not the advice rests on GenAI output (VRGA 5.3, 6.2).", "Professional secrecy holds regardless of working method; client information must not be used to train the model or generate output for other users.", "Client consent must be obtained where the tool's handling of information carries risks that contract terms and safeguards cannot address; standard terms are not sufficient.", "Under VRGA 7.6.2 firms should adopt, publish and regularly update GenAI policies and provide staff training scaled to actual use." ], "relatedIncidents": [], "sources": [ { "label": "Advokatsamfundet — Vägledning om användning av generativ AI i advokatverksamhet (PDF)", "url": "https://www.advokatsamfundet.se/globalassets/advokatsamfundet_sv/advokatyrket/vagledning-om-anvandning-av-generativ-ai-i-advokatverksamhet.pdf" }, { "label": "Advokatsamfundet — Allmänna råd om generativa AI-modeller (13 June 2024)", "url": "https://www.advokatsamfundet.se/Nyhetsarkiv/2024/juni/allmanna-rad-om-generativa-ai-modeller-i-advokatverksamhet/" } ], "country": "SE", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-mi-age-of-ai-report", "url": "https://safelegalai.com/regulation/us-state/us-mi-age-of-ai-report", "title": "State Bar of Michigan AI Workgroup: Transforming the Legal Profession in the Age of AI", "body": "State Bar of Michigan, Board of Commissioners AI Workgroup", "bodyShort": "State Bar of Michigan", "jurisdiction": "us-state", "region": "Michigan", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms", "providers" ], "requirements": [ "competence" ], "versions": [ { "version": "Original", "date": "2025-06-11T00:00:00.000Z", "url": "https://www.michbar.org/Portals/0/publications/pdfs/Age_of_AI_Report_June25.pdf", "changes": "Released by the Board of Commissioners AI Workgroup" } ], "summary": "Michigan's bar-level strategic report on AI, released 11 June 2025. It reaffirms MRPC 1.1 competence as the operative standard, examines AI and the unauthorised practice of law under MCL 600.916, and makes six institutional recommendations on member education, staff development and standing committee oversight. It is a living document, not an ethics opinion.", "keyProvisions": [ "Living document, not a binding ethics opinion", "Reaffirms MRPC 1.1 as the operative competence standard for AI use", "Addresses AI's impact on unauthorised practice of law under MCL 600.916", "Six institutional recommendations covering member education, staff development and standing committee oversight" ], "relatedIncidents": [], "sources": [ { "label": "Transforming the Legal Profession in the Age of AI (PDF)", "url": "https://www.michbar.org/Portals/0/publications/pdfs/Age_of_AI_Report_June25.pdf" }, { "label": "SBM announcement", "url": "https://www.michbar.org/News/NewsDetail/State-Bar-of-Michigan-releases-comprehensive-report-on-AI-and-the-legal-profession" } ], "country": "US", "categories": [ "bar-guidance", "competence-duty", "unauthorised-practice" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "int-cepej-ai-assessment-tool", "url": "https://safelegalai.com/regulation/international/int-cepej-ai-assessment-tool", "title": "Assessment Tool for the operationalisation of the European Ethical Charter on the use of AI in judicial systems (CEPEJ(2023)16)", "body": "European Commission for the Efficiency of Justice, Council of Europe", "bodyShort": "CEPEJ", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "court-staff", "judges", "providers" ], "requirements": [ "risk-classification", "record-keeping" ], "versions": [ { "version": "CEPEJ(2023)16rev", "date": "2025-06-05T00:00:00.000Z", "url": "https://rm.coe.int/cepej-2023-16rev-operationalisation-ai-ethical-charter-en/1680b65218", "changes": "Revised edition of the 2023 assessment tool." }, { "version": "CEPEJ(2023)16final", "date": "2023-12-04T00:00:00.000Z", "url": "https://rm.coe.int/cepej-2023-16final-operationalisation-ai-ethical-charter-en/1680adcc9c" } ], "summary": "A self-assessment questionnaire adopted by CEPEJ in December 2023 and revised on 5 June 2025 that lets judicial institutions evaluate an AI tool or service against the 2018 Ethical Charter before and during deployment. It is aimed at public decision-makers procuring or building AI for courts and complements the Council of Europe's HUDERIA risk-assessment methodology.", "keyProvisions": [ "Structured questions for each Charter principle, applied to a specific AI system used in the judicial field.", "Intended for decision-makers in judicial institutions procuring or building court AI.", "May be supplemented by transparency, security and personal-data tools, and by the HUDERIA methodology.", "Linked to the CEPEJ Resource Centre on Cyberjustice and AI, which catalogued 125 systems by end 2024." ], "relatedIncidents": [], "sources": [ { "label": "CEPEJ(2023)16rev (PDF)", "url": "https://rm.coe.int/cepej-2023-16rev-operationalisation-ai-ethical-charter-en/1680b65218" }, { "label": "CEPEJ — webinar page describing the tool", "url": "https://www.coe.int/en/web/cepej/-/ecn-webinar-9/2024-council-of-europe-framework-convention-on-artificial-intelligence-what-does-it-mean-for-the-judiciary-navigating-huderia-and-cepej-ethical-ai-charter-assessment-tool" } ], "country": "INT", "categories": [ "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-ar-admin-order-25-artificial-intelligence", "url": "https://safelegalai.com/regulation/us-state/us-ar-admin-order-25-artificial-intelligence", "title": "Arkansas Supreme Court Administrative Order No. 25: Artificial Intelligence", "body": "Supreme Court of Arkansas", "bodyShort": "Arkansas Supreme Court", "jurisdiction": "us-state", "region": "Arkansas", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "litigants" ], "requirements": [ "confidentiality", "prohibition" ], "versions": [ { "version": "2025 Ark. 117", "date": "2025-06-05T00:00:00.000Z", "url": "https://arcourts.gov/sites/default/files/In_Re_Creation_of_Admin_Order_No_25_AI.pdf", "changes": "Creation of Administrative Order No. 25" } ], "summary": "Arkansas's court-side AI order. It warns that disclosing confidential or sealed client or case information to a generative AI model may breach Administrative Order No. 19, the Rules of Professional Conduct and the Code of Judicial Conduct, and bars court personnel from exposing CourtConnect data to external generative AI tools without Automation Committee approval.", "keyProvisions": [ "Intentional or inadvertent disclosure of confidential or sealed client or case information to a generative AI model may violate Administrative Order No. 19, the ARPC and the Code of Judicial Conduct", "Court personnel with internal CourtConnect access may not intentionally expose court system data to any external generative AI tool", "An exemption exists for court-approved research and analysis projects with prior approval of the Supreme Court Committee on Automation", "The order does not require pre-filing disclosure or certification of AI use" ], "relatedIncidents": [], "sources": [ { "label": "In re Creation of Administrative Order No. 25 (PDF)", "url": "https://arcourts.gov/sites/default/files/In_Re_Creation_of_Admin_Order_No_25_AI.pdf" }, { "label": "Arkansas Advocate report", "url": "https://arkansasadvocate.com/2025/06/09/arkansas-supreme-court-releases-proposed-rule-for-artificial-intelligence/" } ], "country": "US", "categories": [ "judicial-use", "confidentiality-client-data", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-ar-rpc-ai-amendments-2025", "url": "https://safelegalai.com/regulation/us-state/us-ar-rpc-ai-amendments-2025", "title": "In re Amendments to the Arkansas Rules of Professional Conduct (AI amendments), 2025 Ark. 115", "body": "Supreme Court of Arkansas", "bodyShort": "Arkansas Supreme Court", "jurisdiction": "us-state", "region": "Arkansas", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "supervision", "verification" ], "versions": [ { "version": "2025 Ark. 115", "date": "2025-06-05T00:00:00.000Z", "url": "https://arcourts.gov/", "changes": "Amendments to Rule 5.3 and the Preamble" } ], "summary": "Arkansas wrote AI supervision into its conduct rules rather than issuing a bar opinion. Amended Rule 5.3 makes the supervising attorney fully responsible for AI-generated work product, and the amended Preamble confirms AI use remains subject to all existing rules.", "keyProvisions": [ "Amended Rule 5.3 (Nonlawyer Assistance) makes the supervising attorney personally and fully responsible for AI-generated work product", "The amended Preamble confirms generative AI is not prohibited but its use remains fully subject to all existing rules", "Adopted in place of a numbered Arkansas Bar Association ethics opinion" ], "relatedIncidents": [], "sources": [ { "label": "Arkansas Judiciary (official site)", "url": "https://arcourts.gov/" } ], "country": "US", "categories": [ "bar-guidance", "supervision", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "jp-ai-promotion-act-2025", "url": "https://safelegalai.com/regulation/other/jp-ai-promotion-act-2025", "title": "Act on the Promotion of Research, Development and Utilisation of AI-Related Technology (Law No. 53 of 2025)", "body": "National Diet of Japan", "bodyShort": "Japan Diet", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "permissive" ], "effectiveDate": "2025-09-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-06-04T00:00:00.000Z", "url": "https://laws.e-gov.go.jp/law/507AC0000000053", "changes": "Promulgated and partly in force 4 June 2025; fully in force 1 September 2025 including the AI Strategy Headquarters." } ], "summary": "Japan's first framework AI statute, promulgated 4 June 2025 and fully in force from 1 September 2025. It establishes the AI Strategy Headquarters and a basic AI plan and sets duties of effort for the state, local governments, businesses and citizens. It contains no risk classification, no penalties, and no provisions specific to courts or legal services.", "keyProvisions": [ "Creates the AI Strategy Headquarters headed by the Prime Minister and requires a basic AI plan.", "Duties of effort on businesses to cooperate with national measures; no penalties.", "Government may investigate and publish information on AI-related harms." ], "relatedIncidents": [], "sources": [ { "label": "Cabinet Office AI Act page (official)", "url": "https://www8.cao.go.jp/cstp/ai/ai_act/ai_act.html" }, { "label": "e-Gov law text", "url": "https://laws.e-gov.go.jp/law/507AC0000000053" } ], "country": "JP", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "br-stj-sls-3596-resolve-juizado-2025", "url": "https://safelegalai.com/regulation/brazil/br-stj-sls-3596-resolve-juizado-2025", "title": "STJ SLS 3.596: refusal to suspend AI petition-drafting platform Resolve Juizado", "body": "Superior Court of Justice (Superior Tribunal de Justica)", "bodyShort": "STJ", "jurisdiction": "brazil", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "providers", "litigants" ], "requirements": [ "permissive" ], "effectiveDate": "2025-06-02T00:00:00.000Z", "versions": [ { "version": "Decision of STJ President Herman Benjamin", "date": "2025-06-02T00:00:00.000Z", "url": "https://www.migalhas.com.br/quentes/431663/stj-ministro-mantem-no-ar-site-que-vende-peticoes-geradas-por-ia", "changes": "Initial decision; the merits were later rejected by the 27th Federal Court of Rio de Janeiro in November 2025." } ], "summary": "The President of the Superior Court of Justice refused OAB/RJ's request to suspend a platform selling AI-drafted small-claims petitions for about twenty reais, finding no prima facie unauthorised practice where the law lets citizens litigate without counsel and calling it paradoxical to bar digital tools from those entitled to represent themselves.", "keyProvisions": [ "No prima facie unauthorised practice of law in AI-assisted drafting for small-claims courts.", "Citizens entitled to self-represent may use AI tools.", "Modest fees distinguished from professional legal fees.", "Merits later decided against OAB/RJ by the 27th Federal Court of Rio de Janeiro." ], "relatedIncidents": [], "sources": [ { "label": "Migalhas report", "url": "https://www.migalhas.com.br/quentes/431663/stj-ministro-mantem-no-ar-site-que-vende-peticoes-geradas-por-ia" }, { "label": "ConJur report", "url": "https://conjur.com.br/2025-jun-02/stj-nega-pedido-da-oab-rj-e-mantem-no-ar-site-que-que-vende-peticoes-feitas-por-ia/" } ], "country": "BR", "categories": [ "unauthorised-practice", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "no-ki-lov-consultation", "url": "https://safelegalai.com/regulation/other/no-ki-lov-consultation", "title": "Høring — utkast til ny lov om kunstig intelligens (implementation of the EU AI Act in Norwegian law)", "body": "Ministry of Digitalisation and Public Governance [Digitaliserings- og forvaltningsdepartementet]", "bodyShort": "Norwegian Ministry of Digitalisation", "jurisdiction": "other", "type": "consultation", "status": "consultation", "appliesTo": [ "all", "providers" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Consultation on a draft KI-lov", "date": "2025-06-01T00:00:00.000Z", "url": "https://www.regjeringen.no/no/aktuelt/lov-om-kunstig-intelligens-i-norge-sendes-na-pa-horing/id3113732/", "changes": "Consultation deadline 30 September 2025; entry into force targeted for late summer 2026. Month-only date: the source gives the month, so the 1st is recorded as the date. The consultation draft was not read in full, so this record is unverified." } ], "summary": "A Norwegian Ministry of Digitalisation consultation on a draft Act on artificial intelligence that would incorporate Regulation (EU) 2024/1689 into Norwegian law through the EEA Agreement. It covers the division of responsibility between supervisory authorities, sanctions and geographic scope. The consultation closed on 30 September 2025.", "keyProvisions": [ "Proposes incorporating the AI Act by reference into a new Norwegian KI-lov.", "Addresses allocation of supervisory responsibility and the use of sanctions.", "Consultation deadline 30 September 2025; entry into force targeted for late summer 2026." ], "relatedIncidents": [], "sources": [ { "label": "Regjeringen — Act on artificial intelligence sent for consultation", "url": "https://www.regjeringen.no/no/aktuelt/lov-om-kunstig-intelligens-i-norge-sendes-na-pa-horing/id3113732/" }, { "label": "Datatilsynet — new AI Act sent for consultation", "url": "https://www.datatilsynet.no/aktuelt/aktuelle-nyheter-2025/ny-lov-om-ki-sendt-pa-horing/" } ], "country": "NO", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "bh-iga-general-policy-ai-use", "url": "https://safelegalai.com/regulation/gulf/bh-iga-general-policy-ai-use", "title": "General Policy for the Use of Artificial Intelligence, Version 1.0", "body": "Information and eGovernment Authority (iGA), Kingdom of Bahrain", "bodyShort": "iGA (Bahrain)", "jurisdiction": "gulf", "type": "policy", "status": "in-force", "appliesTo": [ "all", "court-staff" ], "requirements": [ "supervision", "competence", "record-keeping", "confidentiality" ], "versions": [ { "version": "1.0", "date": "2025-05-20T00:00:00.000Z", "url": "https://www.iga.gov.bh/Media/Publications/National%20Digital%20Policies/General%20Policy%20for%20the%20Use%20of%20AI.pdf", "changes": "First release of the policy." } ], "summary": "Bahrain's whole-of-government AI policy, issued by the Information and eGovernment Authority on 20 May 2025. It applies to all government entities and sets four pillars: legal compliance, adoption and use, awareness and education, and cooperation. It contains no judiciary-, court- or legal-services-specific provision.", "keyProvisions": [ "First principle — Human Decision-Making: AI is a tool to assist humans and the final decision in important matters must remain under human control, especially where individual and societal rights are concerned.", "Pillar 1 requires government entities to comply with the Personal Data Protection Law No. 30 of 2018, the Law on Protection of State Documents and Information, Law No. 60 of 2014 on Cybercrime, the Open Data Policy and the GCC AI Ethics guideline.", "Pillar 2 requires entities to consider AI in projects and procurements, to adopt AI only in coordination with the iGA, to follow iGA standards for launching AI initiatives, and to conduct regular audits and reviews.", "Pillar 2 places responsibility for negative outcomes of AI use on responsible individuals rather than the technology.", "Pillar 3 requires government entities to run AI awareness and training programmes for their employees." ], "relatedIncidents": [], "sources": [ { "label": "Official PDF (iga.gov.bh)", "url": "https://www.iga.gov.bh/Media/Publications/National%20Digital%20Policies/General%20Policy%20for%20the%20Use%20of%20AI.pdf" }, { "label": "Official PDF, 30 July 2025 version (iga.gov.bh)", "url": "https://www.iga.gov.bh/Media/Publications/National%20Digital%20Policies/General%20Policy%20for%20the%20Use%20of%20AI%20-%20%20Final%2030%20Jul%202025.pdf" } ], "country": "BH", "categories": [ "court-own-use", "procurement-vendor", "data-protection", "ai-decision-prohibition", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "dk-lov-467-2025-ai-supplementary-provisions", "url": "https://safelegalai.com/regulation/other/dk-lov-467-2025-ai-supplementary-provisions", "title": "Lov om supplerende bestemmelser til forordningen om kunstig intelligens (LOV nr 467 af 14/05/2025)", "body": "Danish Parliament [Folketinget]", "bodyShort": "Denmark", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "all", "providers", "court-staff", "judges" ], "requirements": [ "risk-classification", "supervision", "prohibition" ], "versions": [ { "version": "LOV nr 467 af 14/05/2025", "date": "2025-05-14T00:00:00.000Z", "url": "https://www.retsinformation.dk/api/pdf/249582", "changes": "In force 2 August 2025; does not apply to the Faroe Islands or Greenland." }, { "version": "2025/1 LSF 111 — broader AI-loven bill", "date": "2026-02-18T00:00:00.000Z", "url": "https://www.retsinformation.dk/api/pdf/255520", "changes": "Bill introduced by the Minister for Digitalisation on 18 February 2026 extending the national framework; adoption not confirmed as of 2026-09-04." } ], "summary": "Denmark's national supplement to the EU AI Act, adopted 14 May 2025 and in force from 2 August 2025. It designates Digitaliseringsstyrelsen, Datatilsynet and Domstolsstyrelsen as national competent authorities, sets market surveillance powers, inspection rights and criminal penalties, and creates a bespoke split of supervision over courts' AI use.", "keyProvisions": [ "§ 2(1): Digitaliseringsstyrelsen, Datatilsynet and Domstolsstyrelsen are national competent authorities under Article 70(1) of the AI Act.", "§ 2(4): Domstolsstyrelsen is market surveillance authority for courts' use of AI systems when the courts are not acting in their capacity as a court.", "§ 2(5): where courts use AI in their judicial capacity, the relevant court decides, with appeal to a higher court within four weeks.", "§§ 5–6: market surveillance authorities may demand information and enter business premises without a court order.", "§ 12: entry into force 2 August 2025." ], "relatedIncidents": [], "sources": [ { "label": "Retsinformation — LOV nr 467 af 14/05/2025 (PDF)", "url": "https://www.retsinformation.dk/api/pdf/249582" }, { "label": "Retsinformation — 2025/1 LSF 111 (PDF)", "url": "https://www.retsinformation.dk/api/pdf/255520" } ], "country": "DK", "categories": [ "ai-act-horizontal", "court-own-use", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eu-commission-ai-literacy-qa", "url": "https://safelegalai.com/regulation/eu/eu-commission-ai-literacy-qa", "title": "AI literacy (Article 4 AI Act): Questions and answers", "body": "European Commission", "bodyShort": "European Commission", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "competence" ], "effectiveDate": "2025-05-01T00:00:00.000Z", "versions": [ { "version": "May 2025", "date": "2025-05-01T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers" } ], "summary": "European Commission questions and answers, published May 2025, on the Article 4 AI literacy duty of the EU AI Act that applied from 2 February 2025. Deployers, including law firms that use AI systems in the EU, must ensure staff have sufficient AI literacy proportionate to their role, the systems used and the people affected.", "keyProvisions": [ "Article 4 applies to providers and deployers of any AI system, not only high-risk systems.", "Literacy measures should be proportionate to staff roles, technical knowledge and context of use.", "No mandatory training format or certification; organisations choose the approach and should document it.", "National market surveillance authorities may consider AI literacy when enforcing other obligations." ], "relatedIncidents": [], "article": "does-eu-ai-act-apply-uk-law-firms", "sources": [ { "label": "European Commission — AI literacy Q&A", "url": "https://digital-strategy.ec.europa.eu/en/faqs/ai-literacy-questions-answers" } ], "country": "EU", "categories": [ "ai-act-horizontal", "training-cle", "competence-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "us-md-msba-ai-task-force-advisory", "url": "https://safelegalai.com/regulation/us-state/us-md-msba-ai-task-force-advisory", "title": "MSBA AI and Legal Technology Task Force: Ethical Considerations for Attorney Use of Generative Artificial Intelligence", "body": "Maryland State Bar Association, AI and Legal Technology Task Force", "bodyShort": "MSBA", "jurisdiction": "us-state", "region": "Maryland", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "supervision" ], "versions": [ { "version": "Original", "date": "2025-05-01T00:00:00.000Z", "url": "https://www.msba.org/site/site/rise/Store/StoreLayouts/Item-Detail.aspx?iProductCode=E-6030-25", "changes": "Published by the MSBA AI and Legal Technology Task Force; the source gives a month only (published May 2025), so the day is recorded as the 1st" } ], "summary": "Maryland's only bar-level AI guidance is an MSBA task force advisory from May 2025. It surveys competence, confidentiality, verification and supervision under the Maryland Attorneys' Rules of Professional Conduct but is not a numbered ethics opinion and was not readable in this research run.", "keyProvisions": [ "Non-binding overview of ethical considerations for attorney use of generative AI under the Maryland Attorneys' Rules of Professional Conduct", "Not a numbered ethics opinion; Maryland has no AGC or Supreme Court AI ethics opinion", "Full text distributed through the MSBA store and not verified in this run" ], "relatedIncidents": [], "sources": [ { "label": "MSBA store listing", "url": "https://www.msba.org/site/site/rise/Store/StoreLayouts/Item-Detail.aspx?iProductCode=E-6030-25" } ], "country": "US", "categories": [ "bar-guidance", "confidentiality-client-data", "fees-billing", "competence-duty", "verification-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "israel-ppa-draft-ai-privacy-guidelines", "url": "https://safelegalai.com/regulation/israel/israel-ppa-draft-ai-privacy-guidelines", "title": "Draft Guidelines on the Application of the Privacy Protection Law to Artificial Intelligence Systems", "body": "Privacy Protection Authority of Israel", "bodyShort": "Israel Privacy Protection Authority", "jurisdiction": "israel", "type": "consultation", "status": "consultation", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "consent", "record-keeping", "risk-classification", "confidentiality" ], "versions": [ { "version": "Draft for public comment", "date": "2025-04-30T00:00:00.000Z", "url": "https://www.gov.il/BlobFolder/rfp/ai_reg/he/ai%20regu_draft.pdf", "changes": "First Israeli regulator guidance on AI. Comments closed 5 June 2025; still in draft as of 4 Sep 2026. Still a draft and not read in full, so this record is unverified." } ], "summary": "Draft guidance explaining how the Privacy Protection Law applies across the AI lifecycle, including training and operation. It requires a lawful basis and informed consent for personal data used in AI, transparency that data is collected via an automated AI system, DPO appointment and impact assessments, and internal policies restricting employee use of external generative AI tools.", "keyProvisions": [ "A lawful basis is required for personal data processing at every stage of an AI system's lifecycle, including model training.", "Informed consent must disclose that data is being collected via an AI-based automated system.", "Controllers must appoint DPOs, run data protection impact assessments and adopt written policies on generative AI use.", "Unauthorised web scraping for AI training without consent is not permitted.", "Data subject rights of access, correction and deletion apply and may require changes to the model." ], "relatedIncidents": [], "sources": [ { "label": "PPA draft AI guidelines (official PDF)", "url": "https://www.gov.il/BlobFolder/rfp/ai_reg/he/ai%20regu_draft.pdf" }, { "label": "Gornitzky analysis of the draft", "url": "https://www.gornitzky.com/privacy-in-artificial-intelligence-systems-guidelines-of-the-israeli-privacy-protection-authority/" } ], "country": "IL", "categories": [ "data-protection", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-ak-bar-ethics-opinion-2025-1", "url": "https://safelegalai.com/regulation/us-state/us-ak-bar-ethics-opinion-2025-1", "title": "Alaska Bar Association Ethics Opinion 2025-1: Generative Artificial Intelligence and the Practice of Law", "body": "Alaska Bar Association Ethics Committee", "bodyShort": "Alaska Bar Association", "jurisdiction": "us-state", "region": "Alaska", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "consent", "supervision" ], "versions": [ { "version": "2025-1", "date": "2025-04-23T00:00:00.000Z", "url": "https://alaskabar.org/wp-content/uploads/2025-1.pdf", "changes": "Approved by the Ethics Committee 3 April 2025; adopted by the Board of Governors 23 April 2025" } ], "summary": "Alaska's generative-AI ethics opinion. Lawyers may use generative AI but must understand it, review all output, protect client confidences through closed systems or anonymisation, disclose the basis of any AI charge within a reasonable time of engagement, avoid billing for time not worked, and supervise staff using AI.", "keyProvisions": [ "A lawyer may use generative AI but must first reasonably understand how it works, its limits and whether it safeguards client confidences", "All AI output must be carefully reviewed before use or submission (ARPC 1.1, 3.3)", "Before inputting client confidences the lawyer must review data-retention, data-sharing and self-learning policies; if unprotected, inputs must be anonymised", "AI costs may be billed only if the client's liability and the basis of computation are disclosed within a reasonable time after the representation begins (ARPC 1.5)", "Time saved by AI may not be billed as if worked manually; fees must remain reasonable and proportionate", "Lawyers must comply where a court requires disclosure of generative AI use", "Supervising lawyers must understand the technology well enough to oversee its use (ARPC 5.1, 5.3)" ], "relatedIncidents": [], "sources": [ { "label": "Opinion 2025-1 (PDF)", "url": "https://alaskabar.org/wp-content/uploads/2025-1.pdf" }, { "label": "Adopted opinions index", "url": "https://alaskabar.org/ethics-discipline/ethics-opinions/adopted-ethics-opinions-chronological/" } ], "country": "US", "categories": [ "bar-guidance", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ar-jujuy-scj-acordada-31-2025", "url": "https://safelegalai.com/regulation/other/ar-jujuy-scj-acordada-31-2025", "title": "Acordada 31/2025: Judicial Innovation Laboratory and Protocol for the use of AI in the Judiciary of Jujuy", "body": "Superior Court of Justice of Jujuy (Superior Tribunal de Justicia de Jujuy)", "bodyShort": "STJ Jujuy", "jurisdiction": "other", "region": "Jujuy", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "record-keeping" ], "effectiveDate": "2025-04-11T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-04-11T00:00:00.000Z", "url": "https://www.justiciajujuy.gov.ar/acordadas/acordada-n-31-2025-protocolo-para-el-uso-de-la-inteligencia-artificial-en-el-poder-judicial/", "changes": "Initial protocol (Annex I) and AI-use authorisation form (Annex II)." } ], "summary": "Jujuy's Superior Court created a Judicial Innovation Laboratory and approved a Protocol for the Use of AI in the Judiciary as Annex I, with an authorisation form for AI use as Annex II, applying to all users within the provincial judiciary. The annex text was not accessible, so the substantive parameters remain unverified.", "keyProvisions": [ "Creates the INLAB judicial innovation laboratory.", "Approves the AI-use protocol set out in Annex I.", "Requires an authorisation form for AI use (Annex II).", "Applies to all users within the Jujuy judiciary." ], "relatedIncidents": [], "sources": [ { "label": "Poder Judicial de Jujuy — Acordada 31/2025", "url": "https://www.justiciajujuy.gov.ar/acordadas/acordada-n-31-2025-protocolo-para-el-uso-de-la-inteligencia-artificial-en-el-poder-judicial/" } ], "country": "AR", "categories": [ "judicial-use", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-viac-note-ai-arbitration", "url": "https://safelegalai.com/regulation/international/int-viac-note-ai-arbitration", "title": "VIAC Note on the Use of Artificial Intelligence in Arbitration Proceedings", "body": "Vienna International Arbitral Centre", "bodyShort": "VIAC", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "confidentiality", "competence", "permissive" ], "versions": [ { "version": "First edition", "date": "2025-04-09T00:00:00.000Z", "url": "https://www.viac.eu/wp-content/uploads/2025/04/VIAC-Note-on-AI-1.pdf" } ], "summary": "A short non-binding note published by VIAC in April 2025 for arbitrations under the Vienna Rules 2021. Six headings cover ethical rules, non-delegation of decision-making, confidentiality, responsible and competent use, management of AI use at the case-management conference, and evidence. Arbitrators may inform parties of the AI tools they plan to use and invite comment.", "keyProvisions": [ "Para 2: arbitrators retain full control over decisions and shall not delegate any decision to an AI tool.", "Para 3: arbitrators and parties must ensure AI tools comply with confidentiality obligations and prevent unauthorised access to case data.", "Para 4.2: arbitrators and parties are responsible for the outputs of the AI tools they use.", "Para 5: AI use, disclosure and cost or time impact may be addressed at the case-management conference or in Procedural Order No. 1.", "Para 6: admissibility, relevance and weight of AI-produced evidence are for the tribunal's discretion." ], "relatedIncidents": [], "sources": [ { "label": "VIAC — Note on AI (PDF)", "url": "https://www.viac.eu/wp-content/uploads/2025/04/VIAC-Note-on-AI-1.pdf" } ], "country": "INT", "categories": [ "arbitration-adr", "ai-decision-prohibition", "confidentiality-client-data", "evidence-admissibility", "competence-duty", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pk-sc-ishfaq-ahmed-v-mushtaq-ahmed-2025", "url": "https://safelegalai.com/regulation/other/pk-sc-ishfaq-ahmed-v-mushtaq-ahmed-2025", "title": "Ishfaq Ahmed v Mushtaq Ahmed, 2025 SCP 112 (PLD 2025 SC 582)", "body": "Supreme Court of Pakistan", "bodyShort": "Supreme Court of Pakistan", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "prohibition" ], "effectiveDate": "2025-04-08T00:00:00.000Z", "versions": [ { "version": "Judgment, C.P.L.A. 1010-L/2022", "date": "2025-04-08T00:00:00.000Z", "url": "https://www.supremecourt.gov.pk/downloads_judgements/c.p._1010_l_2022.pdf", "changes": "Original judgment (Mansoor Ali Shah and Aqeel Ahmed Abbasi JJ); exact date within early April 2025 per the Sindh High Court research-cell digest." } ], "summary": "Supreme Court of Pakistan judgment of April 2025 in a rent dispute in which Justice Syed Mansoor Ali Shah examined AI in the judiciary: AI can assist with research, drafting and case management but cannot replace judicial discretion, must respect the Article 10A fair-trial right, and hallucinated references require human verification.", "keyProvisions": [ "AI is an assistive tool; human judges retain final responsibility and discretion.", "AI hallucination and black-box opacity require human oversight of AI-generated legal content.", "Recommends AI-driven, randomised case allocation to reduce partiality.", "NJPMC and the Law and Justice Commission to develop guidelines for permissible AI use in courts.", "Mediation and ADR cannot be replaced by AI." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Pakistan judgment PDF (returned 403 on 4 September 2026)", "url": "https://www.supremecourt.gov.pk/downloads_judgements/c.p._1010_l_2022.pdf" }, { "label": "Sindh High Court Legal Research Cell fortnightly digest, 1-15 April 2025", "url": "https://lrc.shc.gov.pk/downloads/fortnightly_1stapril2025-15aprfinal.pdf" }, { "label": "BABL AI summary", "url": "https://babl.ai/supreme-court-of-pakistan-urges-cautious-ethical-use-of-ai-in-judiciary/" } ], "country": "PK", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "court-own-use", "arbitration-adr" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ar-santa-fe-csj-circular-25-2025", "url": "https://safelegalai.com/regulation/other/ar-santa-fe-csj-circular-25-2025", "title": "Circular No. 25: Good-practice guide for the use of generative AI in the Judiciary of Santa Fe", "body": "Supreme Court of Justice of Santa Fe (Corte Suprema de Justicia de la Provincia de Santa Fe)", "bodyShort": "CSJ Santa Fe", "jurisdiction": "other", "region": "Santa Fe", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality", "supervision", "disclosure" ], "effectiveDate": "2025-04-04T00:00:00.000Z", "versions": [ { "version": "Original (follows the Acuerdo of 6 March 2025 enabling generative-AI tools and Circular 19)", "date": "2025-04-04T00:00:00.000Z", "url": "https://www.justiciasantafe.gov.ar/index.php/circulares/circular-nro-25-guia-de-buenas-practicas-para-el-uso-de-la-inteligencia-artificial-generativa/" } ], "summary": "Minimum acceptable-use standards for judges, defenders and staff granted generative-AI tools by the Santa Fe Supreme Court: independent verification of all outputs, protection of sensitive case data, AI as a complementary assistant with decisions remaining exclusively the judge's, and transparency about AI-generated content, drawing on UNESCO and Council of Europe frameworks.", "keyProvisions": [ "Users must independently verify every AI output.", "Sensitive case data must not be compromised.", "AI cannot replace judicial reasoning; decisions remain the judge's.", "The origin of AI-generated content should be transparent.", "Standards are subject to periodic review." ], "relatedIncidents": [], "sources": [ { "label": "Poder Judicial de Santa Fe — Circular 25", "url": "https://www.justiciasantafe.gov.ar/index.php/circulares/circular-nro-25-guia-de-buenas-practicas-para-el-uso-de-la-inteligencia-artificial-generativa/" }, { "label": "Circular 19 on generative-AI tools", "url": "https://www.justiciasantafe.gov.ar/index.php/circulares/circular-nro-19-uso-de-herramientas-de-inteligencia-artificial-generativa-en-el-poder-judicial-de-santa-fe/" } ], "country": "AR", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "pl-kirp-ai-recommendations", "url": "https://safelegalai.com/regulation/other/pl-kirp-ai-recommendations", "title": "AI w pracy radcy prawnego — Rekomendacje dotyczące korzystania przez radców prawnych z narzędzi opartych na sztucznej inteligencji", "body": "National Chamber of Legal Advisers [Krajowa Izba Radców Prawnych]", "bodyShort": "KIRP", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "consent", "disclosure", "record-keeping" ], "versions": [ { "version": "Wydanie 1, Warszawa 2025 (ISBN 978-83-972004-4-9)", "date": "2025-04-01T00:00:00.000Z", "url": "https://kirp.pl/wp-content/uploads/2025/04/REKOMENDACJE-DOTYCZACE-KORZYSTANIA-PRZEZ-RADCOW-PRAWNYCH-Z-NARZEDZI-OPARTYCH-NA-SZTUCZNEJ-INTELIGENCJI.pdf", "changes": "First edition, published publicly in May 2025; described by KIRP as a living document to be updated. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "A 22-recommendation handbook published by the Polish National Chamber of Legal Advisers in 2025, prepared with Microsoft and the Sołtysiński Kawecki & Szlęzak law firm, on adopting AI tools in legal practice. It runs from tool selection and contract review through professional secrecy, GDPR and client communication to training and incident management.", "keyProvisions": [ "Recommendations 1–11 form a procurement checklist: purpose, data entering the tool, which contract applies, supplier confidentiality commitments, IP and indemnity terms, GDPR role, security measures, integration and fine-tuning.", "Recommendation 12: review client engagement terms for restrictions on AI use.", "Recommendation 13: identify when the client must be informed, including Article 50 AI Act duties.", "Recommendation 14: identify when client consent is needed, given risks such as hallucination.", "Recommendation 15: assess when AI use should affect the client's fee under fixed-fee versus hourly arrangements.", "Recommendations 16–21: internal AI policies, training under Article 4 of the AI Act, regular reassessment and prompt management; Recommendation 22 covers incident management." ], "relatedIncidents": [], "sources": [ { "label": "KIRP — AI w pracy radcy prawnego (PDF)", "url": "https://kirp.pl/wp-content/uploads/2025/04/REKOMENDACJE-DOTYCZACE-KORZYSTANIA-PRZEZ-RADCOW-PRAWNYCH-Z-NARZEDZI-OPARTYCH-NA-SZTUCZNEJ-INTELIGENCJI.pdf" }, { "label": "KIRP — recommendations landing page", "url": "https://kirp.pl/rekomendacje-dotyczace-korzystania-z-ai/" } ], "country": "PL", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "fees-billing", "procurement-vendor", "training-cle", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hk-pcpd-checklist-generative-ai-employees", "url": "https://safelegalai.com/regulation/hong-kong/hk-pcpd-checklist-generative-ai-employees", "title": "Checklist on Guidelines for the Use of Generative AI by Employees", "body": "Office of the Privacy Commissioner for Personal Data, Hong Kong", "bodyShort": "PCPD", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "all" ], "requirements": [ "confidentiality", "verification", "supervision", "record-keeping" ], "effectiveDate": "2025-03-31T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-03-31T00:00:00.000Z", "url": "https://www.pcpd.org.hk/english/news_events/media_statements/press_20250331.html" } ], "summary": "Published 31 March 2025, this PCPD checklist helps organisations, including law firms, draft internal policies on employees' use of generative AI while complying with the Personal Data (Privacy) Ordinance. It covers permitted tools and purposes, what may be entered, verification duties, watermarking, and device and access controls.", "keyProvisions": [ "Specify the permitted generative AI tools, permissible purposes of use, and to whom the policy applies.", "Give clear instructions on the types and amounts of information that may be input, permissible purposes for output, storage and retention.", "Employees are responsible for verifying the accuracy of AI-generated output by proofreading and fact-checking, and for correcting and reporting biased or discriminatory output.", "Give instructions on when and how to watermark or label AI-generated output.", "Specify which devices may be used to access generative AI tools and which categories of employee may use them." ], "relatedIncidents": [], "sources": [ { "label": "PCPD — media statement publishing the checklist (31 Mar 2025)", "url": "https://www.pcpd.org.hk/english/news_events/media_statements/press_20250331.html" } ], "country": "HK", "categories": [ "data-protection", "confidentiality-client-data", "supervision", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "sk-sak-uznesenie-12-4-2025-ai-rules", "url": "https://safelegalai.com/regulation/other/sk-sak-uznesenie-12-4-2025-ai-rules", "title": "Rules on the use of AI tools in advocacy practice (SAK Presidium Resolution 12/4/2025, as amended by 21/14/2025)", "body": "Slovak Bar Association, Presidium [Predsedníctvo Slovenskej advokátskej komory, SAK]", "bodyShort": "SAK", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "consent", "verification", "competence", "supervision", "record-keeping", "prohibition" ], "versions": [ { "version": "12/4/2025", "date": "2025-03-31T00:00:00.000Z", "url": "https://sak.sk/wp-content/uploads/2026/06/Uznesenie_12-4-2025_plne-znenie-k-25-11-2025-ff2a4b92.pdf", "changes": "Original resolution adopted under § 75(1) of Act No. 586/2003 Coll. on Advocacy; effective 1 May 2025." }, { "version": "21/14/2025 consolidated", "date": "2025-11-25T00:00:00.000Z", "url": "https://sak.sk/wp-content/uploads/2026/06/Uznesenie_12-4-2025_plne-znenie-k-25-11-2025-ff2a4b92.pdf", "changes": "Amending resolution of 6 Nov 2025 adding a ten-article methodological guidance and model client clauses; published in Vestník SAK issue 87 and effective 25 Nov 2025." } ], "summary": "Binding Slovak Bar rules on AI in legal practice. Confidential information and personal data may enter AI tools only in one of three modes: a local on-premise tool, a GDPR Art. 28 processor with a no-training guarantee, or express informed client consent waiving confidentiality. Verification of output is mandatory; billing self-training time is barred.", "keyProvisions": [ "Art. 3: AI use must not compromise independence, confidentiality, professional secrecy or conflict rules; the lawyer retains full responsibility for outputs used; AI does not replace professional judgement or legal argument.", "Art. 4: confidential information and personal data must not be entered into AI tools except in mode (a) local tool under the lawyer's exclusive control, (b) express informed client consent, or (c) a GDPR Art. 28 processor with contractual confidentiality.", "Guidance Art. 3(3): relying on AI output without human review is impermissible; the lawyer must actively verify facts, citations, legislation and legal conclusions to avoid hallucinations. Guidance Art. 3(4): the lawyer must demonstrably train trainees and staff.", "Guidance Art. 6: mode (c) requires cumulative minimums — DPA, chain-wide contractual confidentiality, explicit no-training policy, sub-processor transparency, encryption in transit and at rest, MFA, EEA-preferred data location, retention and deletion policies. Enterprise versions strongly recommended; beta versions barred for protected data.", "Guidance Art. 7: mode (b) consent operates as a waiver of confidentiality under § 23(2) of the Act on Advocacy, for a named provider and specific purpose, after disclosure of data-loss, training-use, third-country-transfer and hallucination risks.", "Guidance Art. 8: clients must be informed of AI use where it may materially affect strategy, key decisions or the fee; a lawyer must not charge the client for time spent training themselves on the tool.", "Guidance Art. 9: firms must adopt internal AI rules; recommended controls include an approved/prohibited tool list, regular team training, an AI-use register logging tool, mode and confirmation of human verification, and updated privacy notices naming AI providers." ], "relatedIncidents": [], "sources": [ { "label": "SAK — Uznesenie P SAK č. 12/4/2025 as amended, consolidated text (Vestník SAK čiastka 87, PDF)", "url": "https://sak.sk/wp-content/uploads/2026/06/Uznesenie_12-4-2025_plne-znenie-k-25-11-2025-ff2a4b92.pdf" } ], "country": "SK", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "client-consent-transparency", "competence-duty", "supervision", "training-cle", "procurement-vendor", "fees-billing" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ke-national-ai-strategy-2025-2030", "url": "https://safelegalai.com/regulation/other/ke-national-ai-strategy-2025-2030", "title": "Kenya Artificial Intelligence Strategy 2025-2030", "body": "Ministry of Information, Communications and the Digital Economy", "bodyShort": "MICDE", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "permissive" ], "versions": [ { "version": "Launched", "date": "2025-03-27T00:00:00.000Z", "url": "https://ict.go.ke/sites/default/files/2025-03/Kenya%20AI%20Strategy%202025%20-%202030.pdf", "changes": "Launched at the Kenyatta International Convention Centre, Nairobi." } ], "summary": "Kenya's non-binding national AI strategy, launched 27 March 2025. It sets three pillars — AI digital infrastructure, data ecosystems and governance, and AI research, innovation and commercialisation — with cross-cutting enablers on ethics, skills, finance and governance. It underpins the risk-based logic later carried into the Artificial Intelligence Bill 2026.", "keyProvisions": [ "Three pillars: AI digital infrastructure; data ecosystems and governance; AI research, innovation and commercialisation.", "Cross-cutting enablers covering ethics and responsible AI, skills, financing and institutional governance.", "Developed with German, EU/GIZ, Canadian IDRC and UK FCDO support; non-binding policy, not legislation." ], "relatedIncidents": [], "sources": [ { "label": "Kenya AI Strategy 2025-2030 (Ministry of ICT official PDF)", "url": "https://ict.go.ke/sites/default/files/2025-03/Kenya%20AI%20Strategy%202025%20-%202030.pdf" }, { "label": "Launch announcement (Ministry of ICT official)", "url": "https://www.ict.go.ke/kenyas-artificial-intelligence-ai-strategy-2025-2030-launched-kicc-nairobi" } ], "country": "KE", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ng-ndpa-gaid-2025", "url": "https://safelegalai.com/regulation/other/ng-ndpa-gaid-2025", "title": "Nigeria Data Protection Act General Application and Implementation Directive (GAID) 2025 — Articles 43 and 44 (Emerging Technologies)", "body": "Nigeria Data Protection Commission", "bodyShort": "NDPC", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "all", "firms", "providers" ], "requirements": [ "risk-classification", "record-keeping", "confidentiality", "consent" ], "versions": [ { "version": "NDPC/NDP ACT-GAID/01/2025", "date": "2025-03-20T00:00:00.000Z", "url": "https://ndpc.gov.ng/wp-content/uploads/2025/07/NDP-ACT-GAID-2025-MARCH-20TH.pdf", "changes": "Issued 20 March 2025 (instrument dated 19 December 2024); effective 19 September 2025; replaces the NDPR 2019 as the operative implementing instrument." } ], "summary": "The binding directive implementing Nigeria's Data Protection Act 2023. Article 43 imposes obligations on any controller or processor deploying artificial intelligence to process personal data — documented technical and organisational parameters, a DPIA filed with the Commission, anonymisation and low-risk testing, disparate-outcome review, and continuous monitoring after deployment.", "keyProvisions": [ "Article 43(1): controllers deploying emerging technologies such as AI must take account of the NDP Act, public policy and the GAID to safeguard data-subject privacy.", "Article 43(2): design parameters must reflect the right not to be subject to solely automated decisions, the right to be forgotten, sensitive-data and child safeguards, cross-border rules, and privacy by design and default.", "Article 43(3)-(4): parameters and the DPIA must be documented and filed with the Commission as part of Compliance Audit Returns; the DPIA must assess disparate outcomes and Data Subjects' Vulnerability Indexes.", "Article 43(4)(b)-(e): assess anonymisation, test in low-risk environments, retool and retest, or discard tools posing unmitigable privacy risk; deploy continuous monitoring once judged safe.", "Article 44(3): controllers must refrain from or cease using AI systems impossible to operate in compliance with international human rights law." ], "relatedIncidents": [], "sources": [ { "label": "GAID 2025 (NDPC official PDF)", "url": "https://ndpc.gov.ng/wp-content/uploads/2025/07/NDP-ACT-GAID-2025-MARCH-20TH.pdf" } ], "country": "NG", "categories": [ "data-protection", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ma-cndp-communique-ai-personal-data", "url": "https://safelegalai.com/regulation/other/ma-cndp-communique-ai-personal-data", "title": "Communiqué de presse — IA et protection des données à caractère personnel", "body": "National Commission for the Control of Personal Data Protection [Commission Nationale de Contrôle de la Protection des Données à Caractère Personnel, CNDP]", "bodyShort": "CNDP (Morocco)", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "confidentiality", "disclosure", "consent" ], "versions": [ { "version": "Press release", "date": "2025-03-18T00:00:00.000Z", "url": "https://www.cndp.ma/wp-content/uploads/2025/03/CNDP-Communique-IA-et-protection-des-donnees-a-caractere-personnel-FR.pdf", "changes": "First CNDP public position on AI processing; announces preparation of a deliberation and opens stakeholder hearings." } ], "summary": "Morocco's data protection authority's first public position on AI, issued in Rabat on 18 March 2025. It confirms that AI processing of personal data is governed by Law 09-08, flags automated decision-making as requiring particular safeguards, and announces that work on a dedicated deliberation on AI processing has begun.", "keyProvisions": [ "Lorsqu'ils utilisent les données à caractère personnel, les traitements d'intelligence artificielle (IA) sont encadrés par la loi 09-08.", "Automated decisions require particular attention to integrity, transparency, fairness and legibility so that citizens always retain access to a route of recourse.", "Work on a CNDP deliberation on AI processing has been initiated, following an international benchmark and consultation with foreign data protection authorities.", "Opens hearings with national and international experts, scientific and professional organisations, institutions and civil society. No deliberation adopted as at 4 September 2026." ], "relatedIncidents": [], "sources": [ { "label": "Official PDF (cndp.ma)", "url": "https://www.cndp.ma/wp-content/uploads/2025/03/CNDP-Communique-IA-et-protection-des-donnees-a-caractere-personnel-FR.pdf" }, { "label": "CNDP — AI and personal data protection page", "url": "https://www.cndp.ma/ia-et-protection-des-donnees-a-caractere-personnel/" } ], "country": "MA", "categories": [ "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-ns-court-of-appeal-ai-notice", "url": "https://safelegalai.com/regulation/canada/ca-ns-court-of-appeal-ai-notice", "title": "AI Use in Appeal Proceedings", "body": "Nova Scotia Court of Appeal", "bodyShort": "NS Court of Appeal", "jurisdiction": "canada", "region": "Nova Scotia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "Original", "date": "2025-03-14T00:00:00.000Z", "url": "https://courts.ns.ca/sites/default/files/courts/Court%20of%20Appeal/NSCA_AI_Use_in_Appeal_Proceedings_March_14_2025.pdf", "changes": "Recorded from the court's notice listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A notice of the Nova Scotia Court of Appeal issued on 14 March 2025 governing the use of artificial intelligence in appeal proceedings. It requires parties and counsel to verify every authority and every piece of analysis derived from AI against authoritative sources before it is included in appeal materials.", "keyProvisions": [ "AI-derived authorities and analysis in appeal materials must be verified against authoritative sources." ], "relatedIncidents": [], "sources": [ { "label": "Nova Scotia Courts — Court of Appeal AI notice (PDF)", "url": "https://courts.ns.ca/sites/default/files/courts/Court%20of%20Appeal/NSCA_AI_Use_in_Appeal_Proceedings_March_14_2025.pdf" } ], "country": "CA", "categories": [ "verification-duty", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "it-tribunale-firenze-2025-chatgpt-citations", "url": "https://safelegalai.com/regulation/other/it-tribunale-firenze-2025-chatgpt-citations", "title": "Tribunale di Firenze, Sezione Imprese, order of 14 March 2025 (ChatGPT-invented Cassazione citations)", "body": "Court of Florence (Tribunale di Firenze)", "bodyShort": "Trib. Firenze", "jurisdiction": "other", "region": "Tuscany", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification", "supervision" ], "versions": [ { "version": "Order 14 March 2025, RG 11053/2024", "date": "2025-03-14T00:00:00.000Z", "url": "https://www.litis.it/2025/04/22/chatgpt-in-tribunale-le-allucinazioni-dellia-imbarazzano-un-avvocato-a-firenze/", "changes": "Reported by Litis.it; official text not obtained." } ], "summary": "First reported Italian hallucination case. In an IP interim appeal the respondent's brief cited Cassazione rulings that did not exist; counsel said a trainee had used ChatGPT without his knowledge. The court rejected an Article 96 c.p.c. reckless-litigation claim because the citations did not affect the defence and no bad faith was shown, but censured the failure to verify.", "keyProvisions": [ "Failure to verify AI-generated citations is censured even where no sanction follows.", "Art. 96 c.p.c. requires bad faith or gross negligence plus concrete effect on the proceedings.", "The signing lawyer answers for a collaborator's undisclosed AI use." ], "relatedIncidents": [], "sources": [ { "label": "Litis.it (secondary)", "url": "https://www.litis.it/2025/04/22/chatgpt-in-tribunale-le-allucinazioni-dellia-imbarazzano-un-avvocato-a-firenze/" } ], "country": "IT", "categories": [ "verification-duty", "supervision", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "brazil-cnj-resolucao-615-2025", "url": "https://safelegalai.com/regulation/brazil/brazil-cnj-resolucao-615-2025", "title": "Resolução CNJ nº 615/2025: Use of Artificial Intelligence in the Judiciary", "body": "Conselho Nacional de Justiça", "bodyShort": "CNJ", "jurisdiction": "brazil", "type": "regulation", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "risk-classification", "supervision", "record-keeping", "permissive" ], "effectiveDate": "2025-07-14T00:00:00.000Z", "versions": [ { "version": "Original (approved 11 March 2025)", "date": "2025-03-11T00:00:00.000Z", "url": "https://atos.cnj.jus.br/atos/detalhar/6001" } ], "summary": "Resolution 615/2025 of Brazil's National Council of Justice, approved 11 March 2025 and effective 14 July 2025, replaces Resolution 332/2020 and governs AI across the Brazilian judiciary. It classifies AI uses by risk, requires mandatory human supervision of AI output, creates a National AI Committee, and permits judges and staff to use large language models within those controls.", "keyProvisions": [ "Establishes risk classification for AI systems used by courts, with high-risk uses subject to stricter controls.", "Mandatory human supervision of all AI-assisted judicial work.", "Creates a National AI Committee within the CNJ to oversee compliance.", "Permits use of generative AI and LLMs by judges and staff subject to the resolution.", "Replaces Resolution 332/2020." ], "relatedIncidents": [], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "CNJ — Resolução nº 615/2025", "url": "https://atos.cnj.jus.br/atos/detalhar/6001" } ], "country": "BR", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "au-ncat-procedural-direction-7-generative-ai", "url": "https://safelegalai.com/regulation/australia/au-ncat-procedural-direction-7-generative-ai", "title": "NCAT Procedural Direction 7: Use of Generative Artificial Intelligence (Gen AI)", "body": "New South Wales Civil and Administrative Tribunal", "bodyShort": "NCAT", "jurisdiction": "australia", "region": "New South Wales", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "prohibition", "verification", "confidentiality", "disclosure" ], "effectiveDate": "2025-04-07T00:00:00.000Z", "versions": [ { "version": "Issued 7 March 2025, commenced 7 April 2025", "date": "2025-03-07T00:00:00.000Z", "url": "https://ncat.nsw.gov.au/content/dam/dcj/ctsd/ncat/documents/procedural-directions/ncat-pd7-use-of-generative-ai.pdf", "changes": "Based on Practice Note SC Gen 23, modified for the tribunal's jurisdictions." } ], "summary": "NCAT's procedural direction extends the New South Wales Supreme Court's generative-AI regime to Australia's busiest tribunal. It bars Gen AI from the content of statements, affidavits, statutory declarations and character references, requires verification that must not itself be done with an AI tool, and restricts what may be entered into any program.", "keyProvisions": [ "Gen AI must not generate the content of statements, affidavits, statutory declarations, character references or other evidentiary material, nor be used to alter, embellish, strengthen, dilute or rephrase a witness's written evidence.", "Information subject to non-disclosure or non-publication orders, statutory publication restrictions, or produced under summons must not be entered into any Gen AI program unless the platform is contractually closed and does not train on the data.", "Permitted uses include chronologies, indexes, witness lists, briefs, document and transcript summaries, and written submissions.", "Where Gen AI assisted in written submissions, parties must verify in the body of the submissions that all citations, authorities and legislative references exist, are accurate and are relevant.", "That verification must not be carried out solely by using a Gen AI tool.", "Leave may be sought in exceptional cases to use Gen AI for an annexure or exhibit, identifying the program, version, whether it is open or closed source, and the benefit." ], "relatedIncidents": [], "sources": [ { "label": "NCAT — Procedural Direction 7 (official PDF)", "url": "https://ncat.nsw.gov.au/content/dam/dcj/ctsd/ncat/documents/procedural-directions/ncat-pd7-use-of-generative-ai.pdf" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "evidence-admissibility", "confidentiality-client-data", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cn-cac-ai-content-labeling-measures-2025", "url": "https://safelegalai.com/regulation/other/cn-cac-ai-content-labeling-measures-2025", "title": "Measures for Labelling AI-Generated and Synthesised Content", "body": "Cyberspace Administration of China with MIIT, MPS and NRTA", "bodyShort": "CAC", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "providers" ], "requirements": [ "disclosure" ], "effectiveDate": "2025-09-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-03-07T00:00:00.000Z", "url": "https://www.cac.gov.cn/2025-03/14/c_1743654684782215.htm", "changes": "Issued 7 March 2025; in force 1 September 2025." } ], "summary": "Regulation in force from 1 September 2025 requiring providers to attach explicit labels and implicit metadata labels identifying provider and generation details to AI-generated text, audio, images and video, and requiring distribution platforms to verify metadata and add notices. Relevant to authentication of AI-generated material in litigation, although not itself an evidence rule.", "keyProvisions": [ "Explicit labels on AI-generated text, audio, images and video.", "Implicit metadata labels carrying provider identification.", "Platforms must check metadata and flag AI-generated content." ], "relatedIncidents": [], "sources": [ { "label": "CAC text (official)", "url": "https://www.cac.gov.cn/2025-03/14/c_1743654684782215.htm" } ], "country": "CN", "categories": [ "evidence-admissibility", "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "au-wa-supreme-court-ai-practice-direction-consultation", "url": "https://safelegalai.com/regulation/australia/au-wa-supreme-court-ai-practice-direction-consultation", "title": "Artificial Intelligence practice direction: Consultation note", "body": "Supreme Court of Western Australia", "bodyShort": "Supreme Court of WA", "jurisdiction": "australia", "region": "Western Australia", "type": "consultation", "status": "superseded", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "permissive" ], "effectiveDate": "2025-03-03T00:00:00.000Z", "versions": [ { "version": "Consultation note", "date": "2025-03-03T00:00:00.000Z", "url": "https://www.supremecourt.wa.gov.au/_files/AI_practice_direction.pdf", "changes": "Consultation on whether to make an AI practice direction; superseded by the court's Guidelines for the use of generative AI." } ], "summary": "A consultation note by which the Supreme Court of Western Australia sought input on whether to make a practice direction on generative AI. It surveys the approaches taken in New Zealand, New South Wales, Queensland, Victoria and the United Kingdom, and characterises the New South Wales regime as prohibitive.", "keyProvisions": [ "Consultation limited to generative AI, expressly excluding e-discovery and predictive analytics tools.", "Summarises comparative approaches, describing NSW Practice Note SC Gen 23 as a prohibitive stance.", "The Law Society of Western Australia responded on 11 April 2025 urging a balanced, principles-based approach." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of WA — consultation note (official PDF)", "url": "https://www.supremecourt.wa.gov.au/_files/AI_practice_direction.pdf" }, { "label": "Law Society of WA — submission (11 Apr 2025, PDF)", "url": "https://lawsocietywa.asn.au/wp-content/uploads/2025/04/2025APR11-Submission-to-SCWA-Generative-Artificial-Intelligence.pdf" } ], "country": "AU", "categories": [ "bar-guidance", "verification-duty", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-aaa-icdr-guidance-arbitrators-ai-tools", "url": "https://safelegalai.com/regulation/international/int-aaa-icdr-guidance-arbitrators-ai-tools", "title": "AAA-ICDR Guidance on Arbitrators' Use of AI Tools", "body": "American Arbitration Association - International Centre for Dispute Resolution", "bodyShort": "AAA-ICDR", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence" ], "versions": [ { "version": "March 2025", "date": "2025-03-01T00:00:00.000Z", "url": "https://www.adr.org/", "changes": "Two-page guidance tied to Canons I, V and VI of the Code of Ethics for Arbitrators in Commercial Disputes; official PDF host returned 404 on 3 September 2026." } ], "summary": "A two-page March 2025 note encouraging AAA-ICDR arbitrators to use AI tools within their ethical codes: verify outputs against primary sources, retain complete control over decision-making under Canon V, disclose generative-AI use when it materially affects the process or reasoning, keep confidential information out of tools without data-protection guarantees under Canon VI, and stay competent.", "keyProvisions": [ "Arbitrators should critically evaluate and verify AI outputs and cross-reference them against primary sources.", "Canon V: arbitrators retain complete control over decision-making; AI must support, not replace, judgment.", "Arbitrators should disclose generative-AI use when it materially affects the process or the reasoning of decisions.", "Do not enter party names or case specifics into tools that do not guarantee data protection." ], "relatedIncidents": [], "sources": [ { "label": "AAA-ICDR guidance (official marketing host; returned 404 on 3 September 2026)", "url": "http://go.adr.org/rs/294-SFS-516/images/2025_AAA-ICDR%20Guidance%20on%20Arbitrators%20Use%20of%20AI%20Tools%20(2).pdf" }, { "label": "Text reproduced as Exhibit 1 in ARIAS-US panel materials (October 2025)", "url": "https://www.arias-us.org/wp-content/uploads/2025/10/Exhibits-to-ARIAS-AI-Rules-Panel.pdf" } ], "country": "INT", "categories": [ "arbitration-adr", "verification-duty", "confidentiality-client-data", "ai-decision-prohibition", "competence-duty", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "lsa-gen-ai-rules-of-engagement", "url": "https://safelegalai.com/regulation/canada/lsa-gen-ai-rules-of-engagement", "title": "Gen AI Rules of Engagement for Canadian Lawyers", "body": "Law Society of Alberta", "bodyShort": "Law Society of Alberta", "jurisdiction": "canada", "region": "Alberta", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "effectiveDate": "2024-01-01T00:00:00.000Z", "versions": [ { "version": "March 2025 update", "date": "2025-03-01T00:00:00.000Z", "url": "https://www.lawsociety.ab.ca/resource-centre/key-resources/professional-conduct/gen-ai-rules-of-engagement-for-canadian-lawyers/", "archiveUrl": "https://web.archive.org/web/20260904133822/https://www.lawsociety.ab.ca/resource-centre/key-resources/professional-conduct/gen-ai-rules-of-engagement-for-canadian-lawyers/", "changes": "Refreshed for newer tools and Canadian court practice directions; original published January 2024." }, { "version": "Original", "date": "2024-01-01T00:00:00.000Z", "url": "https://www.lawsociety.ab.ca/resource-centre/key-resources/professional-conduct/gen-ai-rules-of-engagement-for-canadian-lawyers/" } ], "summary": "A practical resource from the Law Society of Alberta, first published January 2024 and updated March 2025, setting out rules of engagement for Canadian lawyers using generative AI. It is permissive with cautions: verify everything, protect confidentiality and privilege, understand the tool, comply with court disclosure directions, and keep the lawyer accountable for the work.", "keyProvisions": [ "Verify every AI-generated fact, citation and legal proposition before use.", "Do not enter confidential or privileged information into tools without adequate protections.", "Comply with court practice directions that require disclosure of AI use.", "Lawyers, not tools, remain accountable to clients, courts and the regulator." ], "relatedIncidents": [ "reddy-v-saroya" ], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Law Society of Alberta — Gen AI Rules of Engagement", "url": "https://www.lawsociety.ab.ca/resource-centre/key-resources/professional-conduct/gen-ai-rules-of-engagement-for-canadian-lawyers/" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-ns-supreme-court-ai-notice", "url": "https://safelegalai.com/regulation/canada/ca-ns-supreme-court-ai-notice", "title": "Ensuring the Integrity of Court Submissions When Using Generative Artificial Intelligence", "body": "Supreme Court of Nova Scotia", "bodyShort": "NS Supreme Court", "jurisdiction": "canada", "region": "Nova Scotia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "Original", "date": "2025-02-28T00:00:00.000Z", "url": "https://courts.ns.ca/sites/default/files/notices/Oct%202023/NSSC_Artificial_Intelligence_Feb_28_2025.pdf", "changes": "Recorded from the court's notice listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A notice of the Supreme Court of Nova Scotia issued on 28 February 2025 on ensuring the integrity of court submissions when generative AI is used. It requires anyone filing materials to verify authorities and analysis derived from generative AI against authoritative sources before those materials are submitted.", "keyProvisions": [ "Authorities and analysis derived from generative AI must be verified against authoritative sources." ], "relatedIncidents": [], "sources": [ { "label": "Nova Scotia Courts — Supreme Court AI notice (PDF)", "url": "https://courts.ns.ca/sites/default/files/notices/Oct%202023/NSSC_Artificial_Intelligence_Feb_28_2025.pdf" } ], "country": "CA", "categories": [ "verification-duty", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-mi-state-bar-ai-faqs", "url": "https://safelegalai.com/regulation/us-state/us-mi-state-bar-ai-faqs", "title": "State Bar of Michigan: AI FAQs for Attorneys", "body": "State Bar of Michigan, Ethics and Practice Management Resource Center", "bodyShort": "State Bar of Michigan", "jurisdiction": "us-state", "region": "Michigan", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "consent", "supervision", "competence" ], "versions": [ { "version": "Fees section update", "date": "2025-02-11T00:00:00.000Z", "url": "https://www.michbar.org/opinions/ethics/AIFAQs", "changes": "Fees section updated" }, { "version": "Original", "date": "2024-11-18T00:00:00.000Z", "url": "https://www.michbar.org/opinions/ethics/AIFAQs", "changes": "Initial publication" } ], "summary": "Michigan's operative AI guidance is a set of State Bar FAQs rather than a formal RI opinion. They impose verification, client-consent, vendor-security, billing and supervision duties under the MRPC, and adopt Ethics Opinion RI-381 as the vendor due-diligence standard for AI providers.", "keyProvisions": [ "Lawyers must understand AI capabilities and limitations including hallucination risk (MRPC 1.1) and must verify all AI-generated citations and text before filing (MRPC 1.3)", "Client consent is required before inputting protected client information into any AI tool, and the provider's security measures must be evaluated per Ethics Opinion RI-381 (MRPC 1.6(c)(1))", "Where AI shortens a task, a lawyer may bill only the actual time worked", "AI subscription costs are firm overhead and may not be passed to clients absent advance agreement; per-use transactional AI costs may be billed as out-of-pocket expenses if disclosed in the engagement agreement", "A firm may not bill clients for time spent learning to use AI tools", "Managing and supervising lawyers must ensure firm personnel comply with the MRPC when using AI (MRPC 5.1, 5.3)", "Submitting AI-hallucinated citations to a court violates MRPC 3.3", "No blanket duty to disclose AI use to clients unless it implicates confidentiality, a court order, or materially affects the representation" ], "relatedIncidents": [], "sources": [ { "label": "State Bar of Michigan AI FAQs", "url": "https://www.michbar.org/opinions/ethics/AIFAQs" }, { "label": "SBM Ethics Opinion RI-381", "url": "https://www.michbar.org/opinions/ethics/numbered_opinions/RI-381" } ], "country": "US", "categories": [ "bar-guidance", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "eu-commission-guidelines-prohibited-ai-practices", "url": "https://safelegalai.com/regulation/eu/eu-commission-guidelines-prohibited-ai-practices", "title": "Commission Guidelines on prohibited artificial intelligence practices (Article 5 AI Act)", "body": "European Commission", "bodyShort": "European Commission", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "prohibition" ], "effectiveDate": "2025-02-04T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-02-04T00:00:00.000Z", "url": "https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-prohibited-artificial-intelligence-ai-practices-defined-ai-act" } ], "summary": "Interpretive guidelines from the European Commission, published 4 February 2025, explaining the prohibited AI practices in Article 5 of the EU AI Act, which applied from 2 February 2025. The guidelines are non-binding but set out how the Commission reads the bans on manipulative techniques, social scoring, untargeted facial scraping and emotion recognition in workplaces.", "keyProvisions": [ "Explains each Article 5 prohibition with examples and scope boundaries.", "Non-binding; authoritative interpretation ultimately rests with the Court of Justice.", "Applies to providers and deployers of AI systems placed on the EU market." ], "relatedIncidents": [], "article": "eu-ai-act-milestones-2024-2026", "sources": [ { "label": "European Commission — guidelines on prohibited AI practices", "url": "https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-prohibited-artificial-intelligence-ai-practices-defined-ai-act" } ], "country": "EU", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "be-ovb-obfg-richtlijnen-ai-advocaten", "url": "https://safelegalai.com/regulation/other/be-ovb-obfg-richtlijnen-ai-advocaten", "title": "Richtlijnen voor advocaten voor het gebruik van Artificiële Intelligentie (OVB/Avocats.be guidelines on AI for lawyers)", "body": "Orde van Vlaamse Balies and Ordre des barreaux francophones et germanophone (Avocats.be)", "bodyShort": "OVB / Avocats.be", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "versions": [ { "version": "First version (bilingual NL/FR)", "date": "2025-02-01T00:00:00.000Z", "url": "https://www.ordevanvlaamsebalies.be/ovb_filesystem/ovb/OVB_OBFG_Richtlijnen-AI-Advocaten.pdf", "changes": "Joint guidelines of the two Belgian bar federations, announced 31 January 2025 and dated 1 February 2025." } ], "summary": "Joint guidelines of Belgium's two bar federations, dated 1 February 2025, on lawyers' use of AI. They add no new deontological duties but apply existing ones: AI is neither forbidden nor mandatory, cited sources must be verified, professional secrecy must never enter a tool, and there is no duty to tell clients AI was used.", "keyProvisions": [ "'Het gebruik van AI is niet verboden, noch verplicht' — use is a matter of the lawyer's own freedom and responsibility.", "The lawyer must know the basics of AI and large language models.", "The lawyer verifies the result including the sources cited — that the legislation, doctrine and case law relied on actually exist.", "No documents or information covered by professional secrecy may be entered into an AI tool, unless the lawyer is absolutely certain it runs in a closed environment with adequate safeguards.", "Personal data must be pseudonymised and kept out of prompts; where processing is essential the lawyer is transparent and seeks consent or another legal basis.", "There is no duty to communicate AI use to the client, any more than for other IT applications; chatbot users must be told the system is automated.", "Terms of use must be examined for training, transfer and storage, further processing, localisation, open/closed character, liability and IP/licence conditions." ], "relatedIncidents": [], "sources": [ { "label": "OVB/Avocats.be — Richtlijnen AI (official PDF)", "url": "https://www.ordevanvlaamsebalies.be/ovb_filesystem/ovb/OVB_OBFG_Richtlijnen-AI-Advocaten.pdf" }, { "label": "OVB — knowledge bank page (official)", "url": "https://www.ordevanvlaamsebalies.be/nl/kennisbank/digitalisering/richtlijnen-voor-advocaten-rond-gebruik-van-artificiele-intelligentie" } ], "country": "BE", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "client-consent-transparency", "procurement-vendor", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "kr-judicial-ai-research-group-guidelines-2025", "url": "https://safelegalai.com/regulation/other/kr-judicial-ai-research-group-guidelines-2025", "title": "Guidelines on the Use of Artificial Intelligence in the Judiciary (Judicial AI Research Group)", "body": "Judicial AI Research Group, a judges' study society", "bodyShort": "Korea Judicial AI Research Group", "jurisdiction": "other", "type": "report", "status": "in-force", "appliesTo": [ "judges", "parties" ], "requirements": [ "disclosure", "verification" ], "effectiveDate": "2025-02-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2025-02-01T00:00:00.000Z", "url": "https://www.legaltimes.co.kr/news/articleView.html?idxno=85019", "changes": "Finalised February 2025; a non-binding recommendation of a judges' research group, not court policy." } ], "summary": "Non-binding guidelines finalised in February 2025 by the Judicial AI Research Group, a study society of judges founded in October 2023. They propose that courts may require parties to disclose AI use in submitted documents, including the tool, prompts and verification steps; warn about deepfake evidence; and state that existing procedural law governs admissibility.", "keyProvisions": [ "Courts may ask parties whether AI was used to prepare documents, images, audio or video and request tool, prompt and verification details.", "Existing civil and criminal procedure rules govern admissibility of AI-generated evidence.", "LLM outputs may contain fabricated citations and must be checked." ], "relatedIncidents": [], "sources": [ { "label": "Legal Times report", "url": "https://www.legaltimes.co.kr/news/articleView.html?idxno=85019" } ], "country": "KR", "categories": [ "disclosure-filings", "evidence-admissibility", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "or-state-bar-formal-opinion-2025-205", "url": "https://safelegalai.com/regulation/us-state/or-state-bar-formal-opinion-2025-205", "title": "Oregon State Bar Formal Opinion 2025-205: Artificial Intelligence Tools", "body": "Oregon State Bar", "bodyShort": "Oregon State Bar", "jurisdiction": "us-state", "region": "Oregon", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "supervision" ], "effectiveDate": "2025-02-01T00:00:00.000Z", "versions": [ { "version": "February 2025", "date": "2025-02-01T00:00:00.000Z", "url": "https://www.osbar.org/_docs/ethics/2025-205.pdf" } ], "summary": "Formal Opinion 2025-205 from the Oregon State Bar, issued February 2025, applies the Oregon Rules of Professional Conduct to lawyers' use of AI tools. Lawyers must understand the tool they use, verify its output before relying on it, protect client confidentiality when entering information, and supervise nonlawyer and vendor use. A 2026 companion opinion addresses chatbots and AI agents.", "keyProvisions": [ "Lawyers must understand the capabilities and limits of the AI tools they use.", "AI output must be verified; lawyers remain responsible for accuracy.", "Client confidential information requires safeguards before entry into AI tools.", "Supervisory duties extend to staff and vendors using AI on client matters." ], "relatedIncidents": [], "sources": [ { "label": "Oregon State Bar — Formal Opinion 2025-205 (PDF)", "url": "https://www.osbar.org/_docs/ethics/2025-205.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "tx-professional-ethics-opinion-705", "url": "https://safelegalai.com/regulation/us-state/tx-professional-ethics-opinion-705", "title": "Opinion 705: Generative Artificial Intelligence", "body": "Professional Ethics Committee for the State Bar of Texas", "bodyShort": "State Bar of Texas", "jurisdiction": "us-state", "region": "Texas", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "verification", "supervision" ], "effectiveDate": "2025-02-01T00:00:00.000Z", "versions": [ { "version": "February 2025", "date": "2025-02-01T00:00:00.000Z", "url": "https://www.legalethicstexas.com/resources/opinions/opinion-705/" } ], "summary": "Opinion 705 from the Professional Ethics Committee for the State Bar of Texas, issued February 2025, addresses lawyers' use of generative AI under the Texas Disciplinary Rules of Professional Conduct. Lawyers must understand the tools they use, protect client confidentiality, verify all output before reliance, supervise staff, and charge only for time actually spent.", "keyProvisions": [ "Competence requires a reasonable understanding of the generative AI tool and its risks.", "Confidential client information must be protected; lawyers should review tool terms on data use.", "AI-generated legal research and drafting must be verified for accuracy.", "Lawyers may not bill clients for time not actually expended." ], "relatedIncidents": [], "sources": [ { "label": "Texas Center for Legal Ethics — Opinion 705", "url": "https://www.legalethicstexas.com/resources/opinions/opinion-705/" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "az-acja-1-509-generative-ai", "url": "https://safelegalai.com/regulation/us-state/az-acja-1-509-generative-ai", "title": "Arizona Code of Judicial Administration 1-509: Use of Generative AI Technology and Large Language Models", "body": "Supreme Court of Arizona", "bodyShort": "Arizona Supreme Court", "jurisdiction": "us-state", "region": "Arizona", "type": "court-rule", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "competence", "confidentiality", "verification" ], "effectiveDate": "2024-10-30T00:00:00.000Z", "versions": [ { "version": "Amended (Administrative Order 2025-24)", "date": "2025-01-29T00:00:00.000Z", "url": "https://www.azcourts.gov/Portals/0/0/admcode/pdfcurrentcode/1-509%20Use%20of%20AI%20Tech%20and%20LLMs%2001_2025.pdf", "changes": "Amendments to the original section adopted by Administrative Order 2025-24." }, { "version": "Original (Administrative Order 2024-207)", "date": "2024-10-30T00:00:00.000Z", "url": "https://www.azcourts.gov/Portals/0/0/admcode/pdfcurrentcode/1-509%20Use%20of%20AI%20Tech%20and%20LLMs%2001_2025.pdf" } ], "summary": "Section 1-509 of the Arizona Code of Judicial Administration, adopted by the Supreme Court of Arizona on 30 October 2024 and amended 29 January 2025, governs how the state judicial branch uses generative AI and large language models. It sets approval, confidentiality, verification and accountability requirements for judges and court employees rather than for litigants.", "keyProvisions": [ "Applies to all Arizona judicial branch personnel using generative AI in their work.", "Confidential, sealed and personally identifiable information may not be entered into public tools.", "Users must review and verify AI output; AI cannot substitute for judicial decision-making.", "Courts must approve tools and maintain an inventory of AI use." ], "relatedIncidents": [], "sources": [ { "label": "Arizona Courts — ACJA 1-509 (PDF)", "url": "https://www.azcourts.gov/Portals/0/0/admcode/pdfcurrentcode/1-509%20Use%20of%20AI%20Tech%20and%20LLMs%2001_2025.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "nsw-sc-gen-23-generative-ai", "url": "https://safelegalai.com/regulation/australia/nsw-sc-gen-23-generative-ai", "title": "Practice Note SC Gen 23: Use of Generative Artificial Intelligence", "body": "Supreme Court of New South Wales", "bodyShort": "NSW Supreme Court", "jurisdiction": "australia", "region": "New South Wales", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "prohibition", "disclosure", "verification" ], "effectiveDate": "2025-02-03T00:00:00.000Z", "versions": [ { "version": "Amended (commenced 3 February 2025)", "date": "2025-01-28T00:00:00.000Z", "url": "https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/general/current/PN_SC_Gen_23.pdf", "changes": "Amended before commencement to clarify permitted uses, the scope of the affidavit disclosure requirement, and the leave process for expert reports." }, { "version": "Original", "date": "2024-11-21T00:00:00.000Z", "url": "https://supremecourt.nsw.gov.au/practice-procedure/generative-artificial-intelligence.html" } ], "summary": "Practice Note SC Gen 23 of the Supreme Court of New South Wales, issued 21 November 2024, amended 28 January 2025 and in force from 3 February 2025. It prohibits using generative AI to draft affidavits, witness statements or expert reports without leave, requires affidavits to disclose that AI was not used, and holds practitioners responsible for verifying submissions.", "keyProvisions": [ "Generative AI must not be used to generate the content of affidavits, witness statements or character references.", "Affidavits and witness statements must contain a disclosure that generative AI was not used.", "Expert reports may not use generative AI without leave of the court.", "Generative AI may be used for submissions and summaries provided all citations are verified.", "Practitioners must not enter confidential or suppressed material into public AI tools." ], "relatedIncidents": [ "valu-v-minister-immigration", "dayal-fedcfamc2f-1166", "jne24-v-minister-immigration" ], "article": "nsw-sc-gen-23-strictest-ai-court-rule", "sources": [ { "label": "NSW Supreme Court — Practice Note SC Gen 23 (PDF)", "url": "https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/general/current/PN_SC_Gen_23.pdf" }, { "label": "NSW Supreme Court — generative AI page", "url": "https://supremecourt.nsw.gov.au/practice-procedure/generative-artificial-intelligence.html" } ], "country": "AU", "categories": [ "disclosure-filings", "verification-duty", "evidence-admissibility", "confidentiality-client-data", "litigant-in-person" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "au-oaic-guidance-commercially-available-ai-products", "url": "https://safelegalai.com/regulation/australia/au-oaic-guidance-commercially-available-ai-products", "title": "Guidance on privacy and the use of commercially available AI products", "body": "Office of the Australian Information Commissioner", "bodyShort": "OAIC", "jurisdiction": "australia", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "confidentiality", "consent", "competence", "record-keeping" ], "effectiveDate": "2024-10-21T00:00:00.000Z", "versions": [ { "version": "Updated", "date": "2025-01-17T00:00:00.000Z", "url": "https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/guidance-on-privacy-and-the-use-of-commercially-available-ai-products", "changes": "Update to the October 2024 guidance." }, { "version": "Original", "date": "2024-10-21T00:00:00.000Z", "url": "https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/guidance-on-privacy-and-the-use-of-commercially-available-ai-products" } ], "summary": "OAIC guidance applying the Privacy Act 1988 and the Australian Privacy Principles to organisations that deploy commercially available AI products. Privacy obligations attach to personal information input into an AI system and to output containing personal information, and the OAIC advises against putting personal information into publicly available generative AI tools.", "keyProvisions": [ "Due diligence is required on AI product selection, covering training data, security and third-party access.", "APP 6 limits use of personal information input into AI systems to the primary purpose of collection absent consent or a reasonable-expectations exception.", "AI-generated or inferred personal information is a collection under APP 3, requiring necessity, lawful and fair means, and consent for sensitive information.", "APP 10 requires reasonable steps to ensure the accuracy of AI-generated personal information.", "Privacy policies must disclose AI use and public-facing AI tools must be identified as such.", "Best practice is not to enter personal information, especially sensitive information, into publicly available generative AI tools." ], "relatedIncidents": [], "sources": [ { "label": "OAIC — guidance on privacy and commercially available AI products", "url": "https://www.oaic.gov.au/privacy/privacy-guidance-for-organisations-and-government-agencies/guidance-on-privacy-and-the-use-of-commercially-available-ai-products" } ], "country": "AU", "categories": [ "data-protection", "procurement-vendor", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "eg-national-ai-strategy-2025-2030", "url": "https://safelegalai.com/regulation/other/eg-national-ai-strategy-2025-2030", "title": "Egypt National Artificial Intelligence Strategy, Second Edition (2025-2030)", "body": "National Council for Artificial Intelligence, Quantum Computing and Emerging Technologies (Arab Republic of Egypt)", "bodyShort": "Egypt NCAI", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "permissive" ], "versions": [ { "version": "Second Edition", "date": "2025-01-16T00:00:00.000Z", "url": "https://ai.gov.eg/SynchedFiles/en/Resources/AIstrategy%20English%2016-1-2025-1.pdf", "changes": "Replaces the first National AI Strategy of May 2021; adds large-model and generative-AI content and sectoral use cases including the judiciary." } ], "summary": "Egypt's second national AI strategy, launched by the President in January 2025 for 2025-2030. It identifies the judiciary as a target sector for the national Arabic large model and records that Egyptian-accent speech recognition has already been used to produce court minutes. It sets no rules for lawyers or judges.", "keyProvisions": [ "Lists judiciary use cases for the national Arabic large model: virtual legal aid guided by applicable laws and judicial precedent; classification of court cases; transcription of legal proceedings; litigation and claim assessment; automated legal research and legal drafting assistance for judges.", "Records as a first-edition achievement the development of Egyptian-accent Arabic speech recognition at 94.1% accuracy, applied to the production of court minutes.", "Pillar 2 (AI for Development) includes AI-based solutions in judicial applications alongside agriculture, health and Arabic NLP.", "Contains no rule, restriction or safeguard governing the use of AI by lawyers, judges or litigants." ], "relatedIncidents": [], "sources": [ { "label": "Official PDF (ai.gov.eg)", "url": "https://ai.gov.eg/SynchedFiles/en/Resources/AIstrategy%20English%2016-1-2025-1.pdf" }, { "label": "OECD.AI policy initiative record", "url": "https://oecd.ai/en/dashboards/policy-initiatives/egypt-national-artificial-intelligence-strategy" } ], "country": "EG", "categories": [ "court-own-use", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hr-pravilnik-anonimizacija-sudskih-odluka-nn-124-24", "url": "https://safelegalai.com/regulation/other/hr-pravilnik-anonimizacija-sudskih-odluka-nn-124-24", "title": "Rulebook on the anonymisation, publication and search of anonymised court decisions (NN 124/24)", "body": "Ministry of Justice, Public Administration and Digital Transformation (Ministarstvo pravosuda, uprave i digitalne transformacije)", "bodyShort": "Ministry of Justice (HR)", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "confidentiality" ], "effectiveDate": "2025-01-01T00:00:00.000Z", "versions": [ { "version": "Narodne novine 124/24; in force 1 January 2025", "date": "2025-01-01T00:00:00.000Z", "url": "https://www.iusinfo.hr/aktualno/u-sredistu/uz-novi-pravilnik-o-anonimiziranju-sudskih-odluka-uvo%C4%91enje-umjetne-inteligencije-i-ponovni-pokusaj-uspostavljanja-ravnoteze-transparentnosti-pravosu%C4%91a-63398", "changes": "Original; official Narodne novine text not opened, secondary analysis cited." } ], "summary": "Implements the 2024 Courts Act duty to publish all court decisions. The ANON AI system performs automatic and semi-automatic anonymisation, metadata extraction and thematic classification; first-instance decisions are anonymised automatically, while higher-court and Supreme Court decisions are anonymised by authorised staff from ANON's proposals. The judge may order further anonymisation where persons could be identified.", "keyProvisions": [ "ANON AI system for anonymisation and indexing of published decisions.", "Human authorised operators finalise anonymisation for higher courts and the Supreme Court.", "Art. 3(3): the judge retains discretion to order further anonymisation." ], "relatedIncidents": [], "sources": [ { "label": "IUS-INFO analysis of the Rulebook (secondary)", "url": "https://www.iusinfo.hr/aktualno/u-sredistu/uz-novi-pravilnik-o-anonimiziranju-sudskih-odluka-uvo%C4%91enje-umjetne-inteligencije-i-ponovni-pokusaj-uspostavljanja-ravnoteze-transparentnosti-pravosu%C4%91a-63398" } ], "country": "HR", "categories": [ "court-own-use", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-co-al-hamim-star-hearthstone", "url": "https://safelegalai.com/regulation/us-state/us-co-al-hamim-star-hearthstone", "title": "Al-Hamim v. Star Hearthstone, LLC, 2024 COA 128 (Colo. Ct. App. 26 Dec 2024)", "body": "Colorado Court of Appeals", "bodyShort": "Colorado Court of Appeals", "jurisdiction": "us-state", "region": "Colorado", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "verification" ], "versions": [ { "version": "2024 COA 128", "date": "2024-12-26T00:00:00.000Z", "url": "https://www.coloradojudicial.gov/system/files/opinions-2024-12/24CA0190-PD.pdf", "changes": "Published opinion" } ], "summary": "The Colorado Court of Appeals published Al-Hamim to address AI hallucinations in appellate briefing. It held that fictitious AI-generated citations violate C.A.R. 28(a)(7)(B), exercised discretion not to sanction the self-represented appellant, and warned that future AI-hallucinated filings may result in sanctions.", "keyProvisions": [ "Submitting a brief containing AI-generated fictitious citations violates C.A.R. 28(a)(7)(B)", "The division declined to sanction the self-represented litigant but stated it will not look kindly on similar infractions in the future", "First Colorado appellate treatment of the appropriate sanction for generative AI hallucinations" ], "relatedIncidents": [], "sources": [ { "label": "Al-Hamim v. Star Hearthstone, LLC (PDF)", "url": "https://www.coloradojudicial.gov/system/files/opinions-2024-12/24CA0190-PD.pdf" } ], "country": "US", "categories": [ "sanctions-record", "verification-duty", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ph-npc-advisory-2024-04-ai-systems", "url": "https://safelegalai.com/regulation/other/ph-npc-advisory-2024-04-ai-systems", "title": "NPC Advisory No. 2024-04: Application of the Data Privacy Act to AI Systems Processing Personal Data", "body": "National Privacy Commission", "bodyShort": "NPC Philippines", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "consent", "confidentiality", "record-keeping", "risk-classification" ], "effectiveDate": "2024-12-19T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-12-19T00:00:00.000Z", "url": "https://www.privacy.gov.ph/wp-content/uploads/2025/02/Advisory-2024.12.19-Guidelines-on-Artificial-Intelligence-w-SGD.pdf" } ], "summary": "The National Privacy Commission's advisory applying Republic Act 10173, its implementing rules and NPC issuances to AI systems that process personal data across development, training, testing and deployment. It is the Philippines' baseline data-protection instrument for legal-sector AI use by firms and public bodies alike.", "keyProvisions": [ "Applies the Data Privacy Act to AI systems across development, training, testing and deployment.", "Addresses lawful bases, transparency and accountability for AI processing of personal data.", "Companion advisory NPC Advisory No. 2026-01 covers scraping of publicly available personal data." ], "relatedIncidents": [], "sources": [ { "label": "NPC — Advisory No. 2024-04 (official PDF)", "url": "https://www.privacy.gov.ph/wp-content/uploads/2025/02/Advisory-2024.12.19-Guidelines-on-Artificial-Intelligence-w-SGD.pdf" }, { "label": "NPC — advisories index", "url": "https://www.privacy.gov.ph/advisories/" } ], "country": "PH", "categories": [ "data-protection", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "il-supreme-court-ai-policy", "url": "https://safelegalai.com/regulation/us-state/il-supreme-court-ai-policy", "title": "Illinois Supreme Court Policy on Artificial Intelligence", "body": "Supreme Court of Illinois", "bodyShort": "Illinois Supreme Court", "jurisdiction": "us-state", "region": "Illinois", "type": "policy", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "litigants" ], "requirements": [ "permissive", "verification" ], "effectiveDate": "2025-01-01T00:00:00.000Z", "versions": [ { "version": "Original (announced December 2024)", "date": "2024-12-18T00:00:00.000Z", "url": "https://ilcourtsaudio.blob.core.windows.net/antilles-resources/resources/e43964ab-8874-4b7a-be4e-63af019cb6f7/Illinois%20Supreme%20Court%20AI%20Policy.pdf" } ], "summary": "A statewide policy from the Supreme Court of Illinois, effective 1 January 2025, authorising the use of AI by litigants, lawyers, judges and court staff. It deliberately does not require disclosure of AI use in pleadings, relying on existing rules of professional conduct and civil procedure to police accuracy, and warns that AI must not compromise due process or privacy.", "keyProvisions": [ "Use of AI by lawyers and self-represented litigants is permitted and expected to expand.", "No requirement to disclose AI use in pleadings; existing rules on accuracy and candor apply.", "Judges and court staff may use AI but retain full responsibility for decisions and work product.", "AI use must not undermine due process, equal protection or access to justice." ], "relatedIncidents": [], "sources": [ { "label": "Illinois Supreme Court — AI Policy (PDF)", "url": "https://ilcourtsaudio.blob.core.windows.net/antilles-resources/resources/e43964ab-8874-4b7a-be4e-63af019cb6f7/Illinois%20Supreme%20Court%20AI%20Policy.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "dk-danske-advokater-ai-guides", "url": "https://safelegalai.com/regulation/other/dk-danske-advokater-ai-guides", "title": "To nye AI-vejledninger til advokatbranchen (AI Act guidance and AI procurement guidance)", "body": "Danske Advokater", "bodyShort": "Danske Advokater", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "competence", "permissive" ], "versions": [ { "version": "Two guidance documents published", "date": "2024-12-17T00:00:00.000Z", "url": "https://www.danskeadvokater.dk/nyheder/to-nye-ai-vejledninger-til-advokatbranchen", "changes": "Published to members on 17 December 2024; full texts are behind the members-only area. Full texts sit behind the members-only area and were not read, so this record is unverified." } ], "summary": "Two guides published on 17 December 2024 by Danske Advokater, the Danish legal-sector trade association. The first explains the EU AI Act's significance for law firms, including risk categories and provider versus deployer roles. The second is a practical checklist of considerations before buying and deploying AI in a firm.", "keyProvisions": [ "Guide 1 maps the AI Act's definitions, risk categories and obligations onto law-firm scenarios.", "Guide 2 covers confidentiality, IP rights, GDPR, ESG considerations and the rules of professional conduct when procuring AI.", "Both are available only in the members' area of the Danske Advokater website." ], "relatedIncidents": [], "sources": [ { "label": "Danske Advokater — two new AI guides", "url": "https://www.danskeadvokater.dk/nyheder/to-nye-ai-vejledninger-til-advokatbranchen" }, { "label": "Danske Advokater — AI theme page (members)", "url": "https://www.danskeadvokater.dk/for-medlemmer/temaer/kunstig-intelligens-ai" } ], "country": "DK", "categories": [ "bar-guidance", "procurement-vendor", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "eu-edpb-opinion-28-2024-ai-models", "url": "https://safelegalai.com/regulation/eu/eu-edpb-opinion-28-2024-ai-models", "title": "Opinion 28/2024 on certain data protection aspects related to the processing of personal data in the context of AI models", "body": "European Data Protection Board", "bodyShort": "EDPB", "jurisdiction": "eu", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "firms", "all" ], "requirements": [ "confidentiality", "record-keeping", "consent" ], "versions": [ { "version": "Adopted", "date": "2024-12-17T00:00:00.000Z", "url": "https://www.edpb.europa.eu/system/files/2024-12/edpb_opinion_202428_ai-models_en.pdf", "changes": "Original opinion adopted under Article 64(2) GDPR at the request of the Irish supervisory authority." } ], "summary": "An EDPB opinion adopted on 17 December 2024 answering when an AI model can be treated as anonymous, how controllers justify legitimate interest at the development and deployment stages, and what follows if training data was processed unlawfully. It matters to law firms and legal-tech vendors deploying AI on personal data in the EU.", "keyProvisions": [ "Anonymity of an AI model must be assessed case by case; the mere absence of stored personal data is not enough.", "Applies the three-step legitimate interest assessment from Guidelines 1/2024 to AI development and deployment.", "Unlawful processing at the development stage may lead supervisory authorities to order fines, processing limitations, dataset erasure or model retraining." ], "relatedIncidents": [], "sources": [ { "label": "EDPB — Opinion 28/2024 (PDF)", "url": "https://www.edpb.europa.eu/system/files/2024-12/edpb_opinion_202428_ai-models_en.pdf" }, { "label": "EDPB — Opinion 28/2024 landing page", "url": "https://www.edpb.europa.eu/documents/opinion-of-the-board-art-64/opinion-282024-on-certain-data-protection-aspects-related-to_en" } ], "country": "EU", "categories": [ "data-protection", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "co-csj-acuerdo-pcsja24-12243", "url": "https://safelegalai.com/regulation/other/co-csj-acuerdo-pcsja24-12243", "title": "Acuerdo PCSJA24-12243: Guidelines for the respectful, responsible, safe and ethical use of AI in the Judicial Branch", "body": "Superior Council of the Judiciary (Consejo Superior de la Judicatura)", "bodyShort": "CSJ", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "disclosure", "prohibition", "confidentiality", "competence", "supervision", "record-keeping" ], "effectiveDate": "2024-12-16T00:00:00.000Z", "versions": [ { "version": "Original (in force on publication)", "date": "2024-12-16T00:00:00.000Z", "url": "https://www.ramajudicial.gov.co/documents/164295699/174277484/PCSJA24-12243.-USO-Y-APROVECHAMIENTO-DE-IA-EN-LA-RAMA-JUDICIAL.pdf", "changes": "Initial guidelines, issued in compliance with Sentencia T-323/2024." } ], "summary": "Binding guidelines for all Colombian judges and judicial employees. Article 8 bars free commercial chatbots, use of AI to assess evidence, weigh facts or decide, and entry of personal or confidential data into tools that retain inputs; article 9 requires verification of every output; article 10 requires express disclosure in the file of tool, model, date, purpose and prompts.", "keyProvisions": [ "Art. 8(3)-(4): no AI for evidence assessment, fact-finding, value judgments or motivating decisions solely on AI output.", "Art. 8(6): no personal, sensitive or confidential data in tools that retain or retrain on inputs.", "Art. 9: verify the veracity, relevance and sufficiency of results and cited sources.", "Art. 10: disclose in the decision the tool, model, provider, date, purpose, prompts and the passages reproduced.", "Art. 13: Judicial School training plan; art. 14: report AI initiatives and malfunctions to the Digital Transformation Unit." ], "relatedIncidents": [], "sources": [ { "label": "Rama Judicial — Acuerdo PCSJA24-12243 (PDF)", "url": "https://www.ramajudicial.gov.co/documents/164295699/174277484/PCSJA24-12243.-USO-Y-APROVECHAMIENTO-DE-IA-EN-LA-RAMA-JUDICIAL.pdf" }, { "label": "Rama Judicial AI page", "url": "https://www.ramajudicial.gov.co/web/utdi/inteligencia-artificial1" }, { "label": "Council of Europe mirror", "url": "https://rm.coe.int/colombia-guidelines-for-the-use-of-artificial-intelligence-in-the-judi/1680b53483" } ], "country": "CO", "categories": [ "judicial-use", "ai-decision-prohibition", "confidentiality-client-data", "court-own-use", "training-cle", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "br-pl-2338-2023-marco-legal-ia", "url": "https://safelegalai.com/regulation/brazil/br-pl-2338-2023-marco-legal-ia", "title": "PL 2338/2023: Legal Framework for Artificial Intelligence (Marco Legal da IA)", "body": "Federal Senate and Chamber of Deputies (Senado Federal / Camara dos Deputados)", "bodyShort": "Congresso Nacional", "jurisdiction": "brazil", "type": "statute", "status": "proposed", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Approved by the Senate plenary", "date": "2024-12-10T00:00:00.000Z", "url": "https://www25.senado.leg.br/web/atividade/materias/-/materia/157233", "changes": "Senate text adopted; sent to the Chamber of Deputies on 17 March 2025 and pending in its Special Committee." }, { "version": "Presented by Sen. Rodrigo Pacheco", "date": "2023-05-03T00:00:00.000Z", "url": "https://www25.senado.leg.br/web/atividade/materias/-/materia/157233" } ], "summary": "Risk-based AI bill modelled on the EU AI Act: it bans excessive-risk systems, imposes obligations on high-risk systems, creates rights to transparency, explanation and contestation, establishes a National AI Regulation and Governance System under the ANPD, and sets fines up to fifty million reais. It passed the Senate on 10 December 2024 and remains pending.", "keyProvisions": [ "Risk tiers: excessive (prohibited), high, and other.", "Rights of affected persons to information, explanation and human review.", "SIA governance system coordinated by the ANPD.", "Administrative fines up to fifty million reais per infringement." ], "relatedIncidents": [], "sources": [ { "label": "Senado Federal — PL 2338/2023", "url": "https://www25.senado.leg.br/web/atividade/materias/-/materia/157233" }, { "label": "Camara Special Committee on PL 2338/23", "url": "https://www2.camara.leg.br/atividade-legislativa/comissoes/comissoes-temporarias/especiais/57a-legislatura/comissao-especial-sobre-inteligencia-artificial-pl-2338-23" } ], "country": "BR", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "au-joint-statement-ai-legal-practice", "url": "https://safelegalai.com/regulation/australia/au-joint-statement-ai-legal-practice", "title": "Statement on the use of artificial intelligence in Australian legal practice", "body": "Victorian Legal Services Board and Commissioner, Law Society of New South Wales and Legal Practice Board of Western Australia", "bodyShort": "VLSB+C / LS NSW / LPBWA", "jurisdiction": "australia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "verification", "disclosure", "competence" ], "effectiveDate": "2024-12-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-12-06T00:00:00.000Z", "url": "https://lsbc.vic.gov.au/sites/default/files/2024-12/Statement%20on%20the%20use%20of%20AI%20in%20Australian%20legal%20practice_2.pdf" } ], "summary": "A joint statement issued 6 December 2024 by the regulators of Victoria, New South Wales and Western Australia on how the Legal Profession Uniform Law and conduct rules apply to AI. Solicitors must not enter confidential client information into public AI tools, must verify AI output, should disclose AI use to clients where appropriate, and remain responsible for their work.", "keyProvisions": [ "Confidential client information must not be entered into public generative AI tools.", "Practitioners must verify AI output and remain responsible for advice and documents.", "Clients should be told about AI use where it is relevant to the engagement or to costs.", "Firms should have policies and training for responsible AI use." ], "relatedIncidents": [], "article": "australia-law-society-ai-guidance-history", "sources": [ { "label": "VLSB+C — joint statement (PDF)", "url": "https://lsbc.vic.gov.au/sites/default/files/2024-12/Statement%20on%20the%20use%20of%20AI%20in%20Australian%20legal%20practice_2.pdf" }, { "label": "VLSB+C — news release", "url": "https://lsbc.vic.gov.au/news-updates/news/statement-use-artificial-intelligence-australian-legal-practice" } ], "country": "AU", "categories": [ "bar-guidance", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "fees-billing" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ee-advokatuur-tehisaru-juhend", "url": "https://safelegalai.com/regulation/other/ee-advokatuur-tehisaru-juhend", "title": "AI guide for advocates [Tehisaru juhend advokaatidele]", "body": "Estonian Bar Association [Eesti Advokatuur], intellectual property and IT law commission", "bodyShort": "Estonian Bar Association", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence" ], "versions": [ { "version": "1.0", "date": "2024-12-03T00:00:00.000Z", "url": "https://advokatuur.ee/en/media/news/tehisaru-juhend-advokaatidele-muutuva-maailma-uued-voimalused-kasikaes-vastutustundliku-lahenemisega", "changes": "Published 3 December 2024, prepared by the Bar's intellectual property and IT law commission, to help advocates and law-office staff use AI effectively and responsibly. Only the Bar's official announcement is public; the guide text was not read, so this record is unverified." } ], "summary": "The Estonian Bar Association's guide on artificial intelligence for advocates, published on 3 December 2024. It is addressed to advocates and law-office staff and sets out verification, competence and confidentiality expectations on the founding principle that the advocate, never the AI system, provides the legal service.", "keyProvisions": [ "'The attorney always provides legal services, not artificial intelligence' — AI output must be critically evaluated, its accuracy verified and its errors corrected by the advocate.", "Advocates must maintain their own expertise and independent judgment rather than letting AI replace professional skill or erode critical thinking.", "Sensitive client information should be entered only into systems whose security has been verified.", "The guide warns against free public AI services, which may use entered information for product development or share it unexpectedly." ], "relatedIncidents": [], "sources": [ { "label": "Eesti Advokatuur — announcement of the AI guide (3 Dec 2024)", "url": "https://advokatuur.ee/en/media/news/tehisaru-juhend-advokaatidele-muutuva-maailma-uued-voimalused-kasikaes-vastutustundliku-lahenemisega" } ], "country": "EE", "categories": [ "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "bar-guidance", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "de-brak-ki-leitfaden", "url": "https://safelegalai.com/regulation/other/de-brak-ki-leitfaden", "title": "Hinweise zum Einsatz von künstlicher Intelligenz (KI)", "body": "Bundesrechtsanwaltskammer (German Federal Bar)", "bodyShort": "BRAK", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "permissive" ], "versions": [ { "version": "Stand Dezember 2024", "date": "2024-12-01T00:00:00.000Z", "url": "https://www.brak.de/fileadmin/service/publikationen/Handlungshinweise/BRAK_Leitfaden_mit_Hinweisen_zum_KI-Einsatz_Stand_12_2024.pdf", "changes": "First edition, announced by BRAK on 9 January 2025. Authored by Dr Frank Remmertz, chair of the BRAK's RDG committee. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "The German Federal Bar's orientation guide, dated December 2024 and announced on 9 January 2025, on using AI in a law firm consistently with professional law. Eight sections read §§ 43, 43a(2), 43e BRAO and § 5 BORA onto AI tools, set out the AI Act duties that fall on lawyers as deployers, and flag copyright, tax and professional-indemnity risks.", "keyProvisions": [ "§ 43 BRAO and the principle of personal performance require independent review and final control of all AI output; AI may support but not replace legal work.", "§ 43a(2) BRAO confidentiality: prompts should be abstract and documents anonymised; under § 43e BRAO, transmitting client secrets to a public LLM is not 'necessary' and is therefore impermissible.", "§ 43e BRAO requires careful provider selection, a contract in at least text form with the statutory minimum content, immediate termination if safeguards lapse, and comparable secrecy protection for foreign providers.", "Neither BRAO nor BORA currently obliges a lawyer to tell the client that AI was used, though transparency and a contractual clause are recommended.", "Article 4 AI Act applies to law firms as 'Betreiber' from 2 February 2025; staff instructions and training are an essential part of responsible use.", "Law-firm AI tools are normally not high-risk, because Annex III point 8(a) targets judicial authorities rather than the bar." ], "relatedIncidents": [], "sources": [ { "label": "BRAK — Hinweise zum Einsatz von künstlicher Intelligenz (KI) (official PDF)", "url": "https://www.brak.de/fileadmin/service/publikationen/Handlungshinweise/BRAK_Leitfaden_mit_Hinweisen_zum_KI-Einsatz_Stand_12_2024.pdf" }, { "label": "BRAK — announcement, 9 January 2025 (official)", "url": "https://www.brak.de/newsroom/newsletter/nachrichten-aus-berlin/2025/ausgabe-1-2025-v-812025/kuenstliche-intelligenz-in-anwaltskanzleien-brak-veroeffentlicht-leitfaden/" } ], "country": "DE", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "procurement-vendor", "ai-act-horizontal", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ontario-rules-civil-procedure-factum-authenticity-certification", "url": "https://safelegalai.com/regulation/canada/ontario-rules-civil-procedure-factum-authenticity-certification", "title": "Ontario Rules of Civil Procedure Rule 4.06.1(2.1) and (2.2): Certification of authenticity of authorities (O. Reg. 384/24)", "body": "Ontario Civil Rules Committee", "bodyShort": "Ontario Rules of Civil Procedure", "jurisdiction": "canada", "region": "Ontario", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "certification", "verification" ], "effectiveDate": "2024-12-01T00:00:00.000Z", "versions": [ { "version": "O. Reg. 384/24", "date": "2024-12-01T00:00:00.000Z", "url": "https://www.ontario.ca/laws/regulation/r24384" } ], "summary": "An amendment to Ontario's Rules of Civil Procedure, in force 1 December 2024, requiring every factum to include a certification that the person signing it is satisfied as to the authenticity of every authority cited. Authorities drawn from CanLII or official court websites are presumed authentic. The rule targets fabricated citations without naming AI.", "keyProvisions": [ "Every factum must certify that the signer is satisfied of the authenticity of each cited authority.", "Authorities from CanLII, official court sites or recognised reporters are presumed authentic.", "Applies to lawyers and self-represented parties filing facta in Ontario civil courts." ], "relatedIncidents": [ "ko-v-li" ], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Ontario e-Laws — O. Reg. 384/24", "url": "https://www.ontario.ca/laws/regulation/r24384" } ], "country": "CA", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "au-nsw-judges-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/australia/au-nsw-judges-generative-ai-guidelines", "title": "Guidelines for New South Wales Judges in respect of use of Generative AI", "body": "Chief Justice of New South Wales", "bodyShort": "NSW Chief Justice", "jurisdiction": "australia", "region": "New South Wales", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "verification", "confidentiality", "supervision", "disclosure" ], "effectiveDate": "2024-11-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-11-21T00:00:00.000Z", "url": "https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/general/current/PN_Generative_AI_21112024.pdf" } ], "summary": "Issued by the Chief Justice of New South Wales on 21 November 2024 alongside Practice Note SC Gen 23 and developed with the other heads of jurisdiction, these guidelines apply to all courts in New South Wales. They prohibit generative AI in the reasoning and drafting of judgments and list red flags for detecting undisclosed AI use.", "keyProvisions": [ "Judges must not use generative AI in formulating reasons for judgment or in assessing or analysing evidence preparatory to reasons.", "Generative AI must not be used for editing or proofing draft judgments, and no part of a draft judgment may be submitted to a generative AI program.", "Any generative AI research output must be closely scrutinised and verified, and must not substitute for personal research by traditional methods.", "Associates, tipstaves and researchers must disclose generative AI use to the judge and separately verify any output.", "Judges may require litigants and legal representatives to disclose AI use and to give an assurance about the verification process followed.", "Lists red flags of AI-generated content, including non-existent citations, out-of-jurisdiction case law, non-specific repetitive language and foreign spelling." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of NSW — Generative AI Practice Note and Judicial Guidelines (official PDF)", "url": "https://supremecourt.nsw.gov.au/documents/Practice-and-Procedure/Practice-Notes/general/current/PN_Generative_AI_21112024.pdf" } ], "country": "AU", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-az-aisc-generative-ai-best-practices", "url": "https://safelegalai.com/regulation/us-state/us-az-aisc-generative-ai-best-practices", "title": "Arizona Steering Committee on AI and the Courts: Generative AI — Ethical Best Practices for Lawyers and Judges", "body": "Arizona Steering Committee on Artificial Intelligence and the Courts", "bodyShort": "Arizona AI Steering Committee", "jurisdiction": "us-state", "region": "Arizona", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "judges" ], "requirements": [ "verification", "competence", "confidentiality", "disclosure" ], "versions": [ { "version": "Original", "date": "2024-11-14T00:00:00.000Z", "url": "https://www.azcourts.gov/Portals/0/74/AISC%20Ethical%20Best%20Practices%20Guidance_For%20Publication.pdf", "changes": "Issued pursuant to Administrative Order 2024-33" } ], "summary": "Supreme Court-sponsored, non-binding Arizona guidance on generative AI for lawyers and judges. It requires users to understand the technology, check all output for accuracy rather than deferring to AI judgment, avoid entering confidential information when in doubt, refrain from billing time saved by AI, and tell prospective clients when they are dealing with an AI intake program.", "keyProvisions": [ "A lawyer must understand the benefits and risks of using generative AI", "A lawyer must check all work product for accuracy and must not defer to the judgment of generative AI", "When in doubt, do not enter confidential or non-public information into a generative AI tool", "A lawyer billing hourly may not bill clients for time saved by using AI", "Where AI is used for client intake, prospective clients must be told they are communicating with an AI program", "Client disclosure is required where the client would need to know about the AI use to make an informed decision" ], "relatedIncidents": [], "sources": [ { "label": "AISC Ethical Best Practices (PDF)", "url": "https://www.azcourts.gov/Portals/0/74/AISC%20Ethical%20Best%20Practices%20Guidance_For%20Publication.pdf" }, { "label": "Administrative Order 2024-33 (PDF)", "url": "https://www.azbar.org/media/4lynbyiv/ao-2024-33.pdf" } ], "country": "US", "categories": [ "bar-guidance", "judicial-use", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-ms-bar-ethics-opinion-267", "url": "https://safelegalai.com/regulation/us-state/us-ms-bar-ethics-opinion-267", "title": "Mississippi Bar Ethics Opinion No. 267: lawyers' use of generative artificial intelligence", "body": "The Mississippi Bar, Committee on Ethics and Professional Responsibility", "bodyShort": "Mississippi Bar", "jurisdiction": "us-state", "region": "Mississippi", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "consent", "supervision" ], "versions": [ { "version": "Opinion No. 267", "date": "2024-11-14T00:00:00.000Z", "url": "https://www.msbar.org/ethics-discipline/ethics-opinions/formal-opinions/267/", "changes": "Issued by the Committee on Ethics and Professional Responsibility" } ], "summary": "Mississippi's generative-AI ethics opinion is unusually specific about client disclosure. Routine AI use needs no disclosure, but three circumstances trigger mandatory disclosure and informed consent, and the opinion imposes an affirmative, risk-calibrated verification duty plus written agreement before AI costs are billed.", "keyProvisions": [ "Lawyers must maintain a reasonable understanding of the benefits and risks of the generative AI tools they use, or draw on others with that expertise (MRPC 1.1)", "Affirmative duty to verify the accuracy and sufficiency of AI work product, calibrated to risk; bulk contract review may be validated on a representative subset (MRPC 1.1, 3.3)", "Reasonable measures to protect confidential information; avoid inputting identifiable client information where possible; informed consent required where inadvertent disclosure remains a concern (MRPC 1.6)", "Three mandatory client disclosure and informed consent triggers (MRPC 1.4): confidential client information will be provided to a third-party AI provider; the AI tool's cost will be charged to the client; or disclosure is required by court rules", "Billing must not duplicate charges or inflate hours; if AI cost is passed to the client, informed consent and a written billing arrangement are required before billing begins (MRPC 1.5)", "Supervisory attorneys must ensure subordinate lawyers and nonlawyer staff comply with confidentiality and verification duties, with training and written policies (MRPC 5.1, 5.3)", "Does not address affirmative disclosure to tribunals absent a court standing order" ], "relatedIncidents": [], "sources": [ { "label": "Mississippi Bar Ethics Opinion 267 landing page", "url": "https://www.msbar.org/ethics-discipline/ethics-opinions/formal-opinions/267/" } ], "country": "US", "categories": [ "bar-guidance", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "br-oab-recomendacao-001-2024", "url": "https://safelegalai.com/regulation/brazil/br-oab-recomendacao-001-2024", "title": "Recomendacao no 001/2024: Responsible Use of Generative AI in Legal Practice", "body": "Federal Council of the Brazilian Bar (Conselho Federal da Ordem dos Advogados do Brasil)", "bodyShort": "OAB", "jurisdiction": "brazil", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "verification", "supervision", "competence", "consent" ], "effectiveDate": "2024-11-11T00:00:00.000Z", "versions": [ { "version": "Approved by the Conselho Pleno (Proposicao 49.0000.2024.007325-9/COP)", "date": "2024-11-11T00:00:00.000Z", "url": "https://s.oab.org.br/arquivos/2024/11/7160d4fe-9449-4aed-80bc-a2d7ac1f5d2f.pdf" } ], "summary": "The Brazilian Bar's first national guidance on generative AI, approved 11 November 2024 and prepared by its Cybersecurity, AI and Data Protection Observatory. Four pillars: applicable law including the LGPD, Statute and Code of Ethics; confidentiality and privacy; ethical legal practice with human supervision and verification; and prior formal communication to clients about AI use.", "keyProvisions": [ "The lawyer must formalise to the client, before use, the intention to use AI in providing services.", "Professional secrecy may not be compromised; no client data in AI systems without confidentiality and security guarantees.", "AI output must be supervised and verified; responsibility stays with the lawyer.", "Attention to algorithmic bias and to keeping up to date on AI implications." ], "relatedIncidents": [], "sources": [ { "label": "OAB — Recomendacao 001/2024 (PDF)", "url": "https://s.oab.org.br/arquivos/2024/11/7160d4fe-9449-4aed-80bc-a2d7ac1f5d2f.pdf" }, { "label": "OAB news: recommendations approved", "url": "https://www.oab.org.br/noticia/62704/oab-aprova-recomendacoes-para-uso-de-ia-na-pratica-juridica" } ], "country": "BR", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "client-consent-transparency", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "br-anpd-tomada-subsidios-ia-decisoes-automatizadas", "url": "https://safelegalai.com/regulation/brazil/br-anpd-tomada-subsidios-ia-decisoes-automatizadas", "title": "ANPD call for input on Artificial Intelligence and Review of Automated Decisions (LGPD art. 20)", "body": "National Data Protection Authority (Autoridade Nacional de Protecao de Dados)", "bodyShort": "ANPD", "jurisdiction": "brazil", "type": "consultation", "status": "consultation", "appliesTo": [ "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "Tomada de subsidios opened", "date": "2024-11-06T00:00:00.000Z", "url": "https://www.gov.br/participamaisbrasil/tomada-de-subsidios-inteligencia-artificial-e-revisao-de-decisoes-automatizadas", "changes": "Opened 6 November 2024; closed 24 January 2025 with 99 contributions (Nota Tecnica 230/2024/CON1/CGN/ANPD)." } ], "summary": "ANPD consultation running from 6 November 2024 to 24 January 2025 on how AI systems and automated decisions map onto LGPD principles, legal bases and the article 20 right to review, as item 7 of its 2025-26 regulatory agenda. A regulation has not been issued; the consultation frames future rules for legal-sector AI processing personal data.", "keyProvisions": [ "Four blocks: LGPD principles, legal bases, data-subject rights, governance and best practice.", "Focus on the art. 20 right to review of automated decisions.", "99 contributions received and consolidated in ANPD technical notes.", "Regulation still pending as of September 2026." ], "relatedIncidents": [], "sources": [ { "label": "Participa + Brasil consultation page", "url": "https://www.gov.br/participamaisbrasil/tomada-de-subsidios-inteligencia-artificial-e-revisao-de-decisoes-automatizadas" } ], "country": "BR", "categories": [ "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "nc-state-bar-2024-formal-ethics-opinion-1", "url": "https://safelegalai.com/regulation/us-state/nc-state-bar-2024-formal-ethics-opinion-1", "title": "2024 Formal Ethics Opinion 1: Use of Artificial Intelligence in a Law Practice", "body": "North Carolina State Bar", "bodyShort": "North Carolina State Bar", "jurisdiction": "us-state", "region": "North Carolina", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "disclosure", "supervision" ], "effectiveDate": "2024-11-01T00:00:00.000Z", "versions": [ { "version": "Adopted", "date": "2024-11-01T00:00:00.000Z", "url": "https://www.ncbar.gov/for-lawyers/ethics/adopted-opinions/2024-formal-ethics-opinion-1/" } ], "summary": "North Carolina State Bar 2024 Formal Ethics Opinion 1, adopted 1 November 2024, permits lawyers in North Carolina to use AI in their practice while applying the Rules of Professional Conduct. Lawyers must be competent in the tools, verify output, protect confidential information, supervise staff, and are encouraged to disclose substantive AI use to clients.", "keyProvisions": [ "Lawyers may use AI tools provided they understand the technology and its limitations.", "Disclosure to clients is recommended where AI is used substantively in the representation.", "All AI-generated research and drafting must be verified for accuracy.", "Client confidentiality must be protected in the selection and use of tools." ], "relatedIncidents": [], "sources": [ { "label": "North Carolina State Bar — 2024 Formal Ethics Opinion 1", "url": "https://www.ncbar.gov/for-lawyers/ethics/adopted-opinions/2024-formal-ethics-opinion-1/" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "ca-qc-barreau-generative-ai-practical-guide", "url": "https://safelegalai.com/regulation/canada/ca-qc-barreau-generative-ai-practical-guide", "title": "L'intelligence artificielle générative: Guide pratique pour une utilisation responsable", "body": "Barreau du Québec", "bodyShort": "Barreau du Québec", "jurisdiction": "canada", "region": "Quebec", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "versions": [ { "version": "Original", "date": "2024-10-25T00:00:00.000Z", "url": "https://www.barreau.qc.ca/media/bnddaqfd/guide-intelligence-artificielle-generative.pdf" } ], "summary": "Launched on 25 October 2024, the Barreau du Quebec's practical guide to responsible use of generative AI applies the Code of Professional Conduct of Lawyers to AI-assisted practice. It requires verification, confidentiality safeguards and competence, and directs members to comply with the AI notices issued by the Quebec and federal courts.", "keyProvisions": [ "Lawyers remain responsible for AI-assisted work under the Code of Professional Conduct of Lawyers.", "Members must comply with the AI notices and directives issued by the Federal Court, Quebec Court of Appeal, Superior Court, Court of Québec and municipal courts.", "Accompanied by member training first delivered on 25 October 2024 and repeated since." ], "relatedIncidents": [], "sources": [ { "label": "Barreau du Québec — guide (PDF)", "url": "https://www.barreau.qc.ca/media/bnddaqfd/guide-intelligence-artificielle-generative.pdf" }, { "label": "Barreau du Québec — regulating AI, work under way", "url": "https://www.barreau.qc.ca/en/new/views-profession/encadrer-ia-multiples-travaux-barreau/" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "de-supreme-court-genai-interim-policy", "url": "https://safelegalai.com/regulation/us-state/de-supreme-court-genai-interim-policy", "title": "Interim Policy on the Use of Generative AI by Judicial Officers and Court Personnel", "body": "Supreme Court of Delaware", "bodyShort": "Delaware Supreme Court", "jurisdiction": "us-state", "region": "Delaware", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "competence", "confidentiality", "verification" ], "effectiveDate": "2024-10-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-10-21T00:00:00.000Z", "url": "https://courts.delaware.gov/forms/download.aspx?id=266868" } ], "summary": "An interim policy adopted by the Supreme Court of Delaware on 21 October 2024 governing how judges and court personnel across the Delaware judiciary may use generative AI. Only approved tools may be used, users must complete training, decision-making cannot be delegated to AI, and users remain responsible for verifying and protecting the confidentiality of any output.", "keyProvisions": [ "Judicial officers and staff may use only generative AI tools approved by the judiciary.", "Training is required before use of an approved tool.", "AI may not be used to make decisions; the human user remains fully responsible.", "Confidential and non-public information must not be entered into unapproved tools." ], "relatedIncidents": [], "sources": [ { "label": "Delaware Courts — Interim GenAI policy (PDF)", "url": "https://courts.delaware.gov/forms/download.aspx?id=266868" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "int-scc-guide-ai-arbitration", "url": "https://safelegalai.com/regulation/international/int-scc-guide-ai-arbitration", "title": "SCC Guide to the use of artificial intelligence in cases administered under the SCC Rules", "body": "SCC Arbitration Institute (Stockholm Chamber of Commerce)", "bodyShort": "SCC", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "parties", "lawyers" ], "requirements": [ "confidentiality", "verification", "disclosure", "permissive" ], "effectiveDate": "2024-10-16T00:00:00.000Z", "versions": [ { "version": "Adopted by the SCC Board", "date": "2024-10-16T00:00:00.000Z", "url": "https://sccarbitrationinstitute.se/wp-content/uploads/2024/12/scc_guide_to_the_use_of_artificial_intelligence_in_cases_administered_under_the_scc_rules-1.pdf", "changes": "Announced 17 October 2024." } ], "summary": "Adopted by the SCC Board on 16 October 2024, this light-touch guide offers flexible, non-mandatory guidance to tribunals and participants in SCC cases. Citing the EU AI Act's high-risk classification of AI used by tribunals, it addresses confidentiality, quality through human review of outputs, integrity through disclosure of AI use, and non-delegation of arbitral decision-making.", "keyProvisions": [ "Confidentiality: participants, especially tribunals, should understand how input data is used by AI tools.", "Quality: tribunals should apply an appropriate level of review and verification to AI outputs before use.", "Integrity: tribunals are encouraged to disclose any use of AI in researching or interpreting facts and law.", "Non-delegation: AI may support but cannot replace arbitral decision-making or its reasoning." ], "relatedIncidents": [], "sources": [ { "label": "SCC — Guide (PDF)", "url": "https://sccarbitrationinstitute.se/wp-content/uploads/2024/12/scc_guide_to_the_use_of_artificial_intelligence_in_cases_administered_under_the_scc_rules-1.pdf" }, { "label": "SCC — announcement, 17 October 2024", "url": "https://sccarbitrationinstitute.se/en/news/the-scc-arbitration-institute-leads-the-way-in-embracing-the-age-of-ai/" } ], "country": "INT", "categories": [ "arbitration-adr", "ai-decision-prohibition", "evidence-admissibility", "procurement-vendor", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-ns-registrar-bankruptcy-ai-directive", "url": "https://safelegalai.com/regulation/canada/ca-ns-registrar-bankruptcy-ai-directive", "title": "Artificial Intelligence Directive, Registrar in Bankruptcy", "body": "Registrar in Bankruptcy, Supreme Court of Nova Scotia", "bodyShort": "NS Registrar in Bankruptcy", "jurisdiction": "canada", "region": "Nova Scotia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure" ], "versions": [ { "version": "Original", "date": "2024-10-04T00:00:00.000Z", "url": "https://courts.ns.ca/sites/default/files/notices/2024-10/AI_Directive_Registrar_in_Bankruptcy_Oct_4_2024.pdf", "changes": "Recorded from the court's notice listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A directive of the Registrar in Bankruptcy in Nova Scotia issued on 4 October 2024, extending the Nova Scotia courts' approach to artificial intelligence to bankruptcy proceedings. It requires disclosure of AI use in materials filed with the Registrar and places responsibility for their accuracy on the filing party.", "keyProvisions": [ "AI use in materials filed in bankruptcy proceedings must be disclosed." ], "relatedIncidents": [], "sources": [ { "label": "Nova Scotia Courts — Registrar in Bankruptcy AI directive (PDF)", "url": "https://courts.ns.ca/sites/default/files/notices/2024-10/AI_Directive_Registrar_in_Bankruptcy_Oct_4_2024.pdf" } ], "country": "CA", "categories": [ "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-il-22nd-circuit-ao-2024-23", "url": "https://safelegalai.com/regulation/us-state/us-il-22nd-circuit-ao-2024-23", "title": "Illinois 22nd Judicial Circuit (McHenry County) Administrative Order 2024-23: Artificial Intelligence Policy and Plan", "body": "Twenty-Second Judicial Circuit of Illinois (McHenry County)", "bodyShort": "IL 22nd Judicial Circuit", "jurisdiction": "us-state", "region": "Illinois", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "judges", "court-staff", "litigants" ], "requirements": [ "disclosure", "verification" ], "versions": [ { "version": "AO 2024-23", "date": "2024-10-01T00:00:00.000Z", "url": "https://www.22ndcircuitil.gov/wp-content/uploads/sites/3/2024/10/Administrative-Order-2024-23-Artificial-Intelligence-Policy-and-Plan.pdf", "changes": "Initial administrative order; the source gives a month only (issued October 2024), so the day is recorded as the 1st" } ], "summary": "The 22nd Judicial Circuit of Illinois adopted a circuit-level AI policy and plan by administrative order in October 2024, ahead of and going beyond the Illinois Supreme Court's statewide policy. Its operative terms were not read in this research run and must be confirmed.", "keyProvisions": [ "Circuit-level artificial intelligence policy and plan for McHenry County courts (terms not yet verified)" ], "relatedIncidents": [], "sources": [ { "label": "Administrative Order 2024-23 (PDF)", "url": "https://www.22ndcircuitil.gov/wp-content/uploads/sites/3/2024/10/Administrative-Order-2024-23-Artificial-Intelligence-Policy-and-Plan.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "judicial-use", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-nm-ethics-advisory-opinion-2024-004", "url": "https://safelegalai.com/regulation/us-state/us-nm-ethics-advisory-opinion-2024-004", "title": "Formal Ethics Advisory Opinion 2024-004: Using Generative Artificial Intelligence in the Practice of Law", "body": "State Bar of New Mexico Ethics Advisory Committee", "bodyShort": "State Bar of New Mexico", "jurisdiction": "us-state", "region": "New Mexico", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "consent", "supervision" ], "versions": [ { "version": "Original", "date": "2024-09-24T00:00:00.000Z", "url": "https://www.sbnm.org/Portals/NMBAR/GenAI%20Formal%20Opinion%20-%20Sept_2024_FINAL.pdf", "changes": "Initial publication" } ], "summary": "Formal Opinion 2024-004 answers yes, qualified, to whether New Mexico lawyers may use generative AI. It implicates Rules 16-101, 16-103, 16-104, 16-105, 16-106, 16-109, 16-303, 16-501, 16-503 and 16-804(C) NMRA, and requires competence, independent verification of output, confidentiality safeguards, conflicts checking and reasonable fees. The opinion is advisory and not binding on tribunals.", "keyProvisions": [ "Responsible generative AI use is consistent with the duties of competence, diligence and charging reasonable fees", "Lawyers must not input confidential information into tools lacking adequate confidentiality and security protections, and should obtain informed client consent before doing so", "Before filing, a lawyer must independently verify the accuracy and sufficiency of all AI-generated material (Rules 16-303, 16-804(C))", "Lawyers must check court, agency and local rules on whether AI use is allowed and what disclosure is required", "A lawyer may bill only for time actually spent when AI reduces the work, under Rule 16-105", "Prior-client information retained by an AI tool can create Rule 16-109 conflicts that must be independently ruled out" ], "relatedIncidents": [], "sources": [ { "label": "State Bar of New Mexico Formal Ethics Advisory Opinion 2024-004 (official PDF)", "url": "https://www.sbnm.org/Portals/NMBAR/GenAI%20Formal%20Opinion%20-%20Sept_2024_FINAL.pdf" } ], "country": "US", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "fees-billing", "competence-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "sg-registrar-circular-1-2024-generative-ai", "url": "https://safelegalai.com/regulation/singapore/sg-registrar-circular-1-2024-generative-ai", "title": "Registrar's Circular No. 1 of 2024: Guide on the Use of Generative AI Tools by Court Users", "body": "Supreme Court of Singapore (with State Courts and Family Justice Courts)", "bodyShort": "Singapore Courts", "jurisdiction": "singapore", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "permissive" ], "effectiveDate": "2024-10-01T00:00:00.000Z", "versions": [ { "version": "Original (issued 23 September 2024)", "date": "2024-09-23T00:00:00.000Z", "url": "https://www.judiciary.gov.sg/docs/default-source/news-and-resources-docs/guide-on-the-use-of-generative-ai-tools-by-court-users.pdf" } ], "summary": "A guide issued by the Singapore Supreme Court on 23 September 2024 and effective 1 October 2024, mirrored by the State Courts and Family Justice Courts. Court users may use generative AI but are fully responsible for the content, must verify it, must not fabricate evidence, need not declare AI use unless asked, and face personal costs for fabricated authorities.", "keyProvisions": [ "No pre-emptive declaration of AI use is required unless the court asks.", "The user is responsible for all content filed, including verifying citations.", "Generative AI must not be used to fabricate or alter evidence.", "Personal costs orders may follow the citation of non-existent authorities." ], "relatedIncidents": [ "tan-hai-peng-v-tan-cheong-joo" ], "article": "singapore-ai-guidance-history", "sources": [ { "label": "Singapore Courts — Guide on the use of generative AI tools by court users (PDF)", "url": "https://www.judiciary.gov.sg/docs/default-source/news-and-resources-docs/guide-on-the-use-of-generative-ai-tools-by-court-users.pdf" }, { "label": "Singapore Courts — Supreme Court registrar's circulars", "url": "https://www.judiciary.gov.sg/news-and-resources/registrar's-circulars/Courts/supreme-court" } ], "country": "SG", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person", "evidence-admissibility", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-iba-caidp-future-is-now-report", "url": "https://safelegalai.com/regulation/international/int-iba-caidp-future-is-now-report", "title": "The Future is Now: Artificial Intelligence and the Legal Profession", "body": "International Bar Association with the Center for AI and Digital Policy", "bodyShort": "IBA/CAIDP", "jurisdiction": "international", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "supervision", "disclosure" ], "versions": [ { "version": "September 2024", "date": "2024-09-19T00:00:00.000Z", "url": "https://www.ibanet.org/document?id=The-future-is%20now-AI-and-the-legal-profession-report", "changes": "Launched at the IBA Annual Conference, Mexico City, 19 September 2024; released 30 September 2024." } ], "summary": "A September 2024 IBA and CAIDP report on AI's impact on law firms, the profession and regulation, based on a survey of 210 firms of which 43 per cent had no AI policy. It recommends AI governance policies, AI-literacy training, updated ethics rules covering supervision of AI tools and disclosure of AI use, and consistent cross-border regulation.", "keyProvisions": [ "Recommends firms adopt AI policies covering data governance, security, intellectual property and privacy.", "Calls for ethics guidelines to cover proper supervision of AI tools and disclosure obligations regarding AI use.", "Recommends AI-literacy training programmes for lawyers, with support for smaller firms.", "Urges consistent, coherent AI regulation with bars as key stakeholders." ], "relatedIncidents": [], "sources": [ { "label": "IBA — report page", "url": "https://www.ibanet.org/IBA-and-CAIDP-release-report-The-Future-is-Now-Artificial-Intelligence-and-the-Legal-Profession" }, { "label": "IBA — report (PDF)", "url": "https://www.ibanet.org/document?id=The-future-is%20now-AI-and-the-legal-profession-report" } ], "country": "INT", "categories": [ "bar-guidance", "competence-duty", "supervision", "client-consent-transparency", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ar-aaip-guia-ia-responsable-2024", "url": "https://safelegalai.com/regulation/other/ar-aaip-guia-ia-responsable-2024", "title": "Guide for public and private entities on transparency and personal-data protection for responsible AI", "body": "Agency for Access to Public Information (Agencia de Acceso a la Informacion Publica)", "bodyShort": "AAIP", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification", "record-keeping" ], "versions": [ { "version": "Published under the programme created by Resolucion 161/2023", "date": "2024-09-01T00:00:00.000Z", "url": "https://www.argentina.gob.ar/noticias/guia-de-la-aaip-para-usar-la-inteligencia-artificial-de-manera-responsable", "changes": "Initial guide; a preliminary version circulated earlier in 2024." } ], "summary": "Argentina's data-protection authority's lifecycle guide for AI systems: privacy by design and impact assessment at design stage, bias and error testing at validation, documented privacy policies at implementation, and incident reporting, monitoring and complaint channels in operation. Not legal-sector specific but applies to firms and courts processing personal data with AI.", "keyProvisions": [ "Data-protection impact evaluation from the design stage.", "Bias detection and ethical-commitment documentation before deployment.", "Transparency to data subjects and traceability of systems.", "Incident reporting and accessible complaint channels in operation." ], "relatedIncidents": [], "sources": [ { "label": "Argentina.gob.ar news", "url": "https://www.argentina.gob.ar/noticias/guia-de-la-aaip-para-usar-la-inteligencia-artificial-de-manera-responsable" }, { "label": "IAPP report", "url": "https://iapp.org/news/a/la-autoridad-de-protecci-n-de-datos-personales-de-argentina-publica-su-gu-a-para-el-uso-responsable-de-la-inteligencia-artificial" } ], "country": "AR", "categories": [ "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ca-cjc-ai-guidelines-canadian-courts", "url": "https://safelegalai.com/regulation/canada/ca-cjc-ai-guidelines-canadian-courts", "title": "Guidelines for the Use of Artificial Intelligence in Canadian Courts", "body": "Canadian Judicial Council [Conseil canadien de la magistrature]", "bodyShort": "CJC", "jurisdiction": "canada", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "supervision", "confidentiality", "competence", "prohibition" ], "versions": [ { "version": "September 2024", "date": "2024-09-01T00:00:00.000Z", "url": "https://cjc-ccm.ca/sites/default/files/documents/2024/AI%20Guidelines%20-%20FINAL%20-%202024-09%20-%20EN.pdf", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the CJC publication page; the guidelines PDF was not read in full, so this record is unverified." } ], "summary": "Guidelines drafted by the Canadian Judicial Council's Technology Committee and published in September 2024 for federally appointed judges. They raise awareness of AI risks and set a principled framework for appropriate use: judicial independence, ethical use, safety, transparency, monitoring and continuing education. Judges retain exclusive responsibility for their decisions and may not delegate decision-making.", "keyProvisions": [ "Judges hold exclusive responsibility for their decisions and are not permitted to delegate decision-making authority to AI.", "Any AI use must preserve judicial independence and comply with the Ethical Principles for Judges.", "Transparency, safety, monitoring of AI tools and continuing judicial education are required.", "Published in both official languages and adopted or endorsed by several provincial courts." ], "relatedIncidents": [], "sources": [ { "label": "CJC — Guidelines for the Use of AI in Canadian Courts (PDF)", "url": "https://cjc-ccm.ca/sites/default/files/documents/2024/AI%20Guidelines%20-%20FINAL%20-%202024-09%20-%20EN.pdf" }, { "label": "CJC — publication page", "url": "https://cjc-ccm.ca/en/resources-center/publications/guidelines-use-artificial-intelligence-canadian-courts" } ], "country": "CA", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "fr-cnb-guide-pratique-ia-generative", "url": "https://safelegalai.com/regulation/other/fr-cnb-guide-pratique-ia-generative", "title": "Guide pratique — Utilisation des systèmes d'intelligence artificielle générative", "body": "Conseil national des barreaux (National Bar Council), France", "bodyShort": "CNB", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "competence", "verification" ], "versions": [ { "version": "1ère édition", "date": "2024-09-01T00:00:00.000Z", "url": "https://cnb.avocat.fr/medias/cnb-guidepratique-utilisation-systemes-iag-2024-68f7b1d86e4570.98487114.pdf", "changes": "First edition, produced by the CNB's Groupe de travail Intelligence Artificielle. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "A 42-page practical guide published by the CNB's AI working group in September 2024 to help French lawyers understand and adopt generative AI. It explains how generative AI is built, sets out use cases and risks, and gives guidance on formulating prompts, evaluating and improving generated output, and training. It is accompanied by a tool-selection grid.", "keyProvisions": [ "Uses the Article 3(1) definition of an AI system from Regulation (EU) 2024/1689.", "Sets out use cases across correspondence, drafting, contracts, summarising and research, and the associated risks.", "Gives practical method for prompting and for evaluating, improving and using generated output.", "Accompanied by a selection grid scoring tools on data sovereignty, security, functionality, ethical compliance and cost.", "Points lawyers to CNB training on generative AI." ], "relatedIncidents": [], "sources": [ { "label": "CNB — Guide pratique, utilisation des systèmes d'IA générative (official PDF)", "url": "https://cnb.avocat.fr/medias/cnb-guidepratique-utilisation-systemes-iag-2024-68f7b1d86e4570.98487114.pdf" }, { "label": "CNB — Le CNB accompagne toujours plus la profession dans l'utilisation de l'IA générative (official)", "url": "https://cnb.avocat.fr/actualite/le-cnb-accompagne-toujours-plus-la-profession-dans-l-utilisation-de-l-ia-generative" } ], "country": "FR", "categories": [ "bar-guidance", "competence-duty", "verification-duty", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "my-mosti-national-guidelines-ai-governance-ethics", "url": "https://safelegalai.com/regulation/other/my-mosti-national-guidelines-ai-governance-ethics", "title": "National Guidelines on AI Governance and Ethics (AIGE)", "body": "Ministry of Science, Technology and Innovation [Kementerian Sains, Teknologi dan Inovasi]", "bodyShort": "MOSTI", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "permissive", "risk-classification" ], "effectiveDate": "2024-09-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-09-01T00:00:00.000Z", "url": "https://www.mosti.gov.my/wp-content/uploads/2024/09/NATIONAL-GUIDELINES-OF-AIGE-20241118.pdf", "changes": "Only the month of publication (September 2024) is established, so the date is recorded as the first of that month; the published file itself carries the date 18 November 2024." } ], "summary": "Malaysia's voluntary national AI governance and ethics guidelines, published by MOSTI in September 2024. They set cross-sector principles for AI developers, deployers and end users, sit alongside the National AI Action Plan, and contain no provisions specific to courts, judges or the legal profession.", "keyProvisions": [ "Voluntary cross-sector AI governance and ethics principles.", "Addressed to AI developers, deployers and end users.", "Contains no provisions specific to courts, judges or legal services." ], "relatedIncidents": [], "sources": [ { "label": "MOSTI — National Guidelines on AI Governance and Ethics (official PDF)", "url": "https://www.mosti.gov.my/wp-content/uploads/2024/09/NATIONAL-GUIDELINES-OF-AIGE-20241118.pdf" } ], "country": "MY", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-mn-olpr-genai-ethics-guidance", "url": "https://safelegalai.com/regulation/us-state/us-mn-olpr-genai-ethics-guidance", "title": "Minnesota Office of Lawyers Professional Responsibility: ethics guidance for generative AI use", "body": "Minnesota Office of Lawyers Professional Responsibility", "bodyShort": "Minnesota OLPR", "jurisdiction": "us-state", "region": "Minnesota", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "consent", "verification", "competence" ], "versions": [ { "version": "Original", "date": "2024-09-01T00:00:00.000Z", "url": "https://www.mnbars.org/?pg=BenchBarofMinnesota&pubAction=viewIssue&pubIssueID=47719&pubIssueItemID=294962", "changes": "Published in Bench and Bar of Minnesota; the source gives a month only (published September 2024), so the day is recorded as the 1st" } ], "summary": "Minnesota's disciplinary regulator issued informal generative-AI guidance through an article by the OLPR Director in September 2024. It is the closest thing Minnesota has to an ethics opinion and is reported to make documented informed consent for AI use of client data an affirmative Rule 1.6 obligation.", "keyProvisions": [ "Informal guidance from the Director of the Office of Lawyers Professional Responsibility, not a numbered LPRB opinion", "Reported to treat documented informed consent for client data used with AI tools as an affirmative obligation under Minn. R. Prof. Conduct 1.6", "Text not opened in this research run and must be confirmed before publication" ], "relatedIncidents": [], "sources": [ { "label": "Bench and Bar of Minnesota, September 2024", "url": "https://www.mnbars.org/?pg=BenchBarofMinnesota&pubAction=viewIssue&pubIssueID=47719&pubIssueItemID=294962" } ], "country": "US", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "competence-duty", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ng-nba-slp-ai-guidelines-legal-profession", "url": "https://safelegalai.com/regulation/other/ng-nba-slp-ai-guidelines-legal-profession", "title": "Guidelines for the Use of Artificial Intelligence in the Legal Profession in Nigeria", "body": "Nigerian Bar Association — Section on Legal Practice, Technology and Law Committee", "bodyShort": "NBA-SLP", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence", "supervision", "consent", "disclosure" ], "versions": [ { "version": "As launched", "date": "2024-08-28T00:00:00.000Z", "url": "https://blog.nigerianbar.org.ng/2024/09/14/nba-launches-new-technology-guidelines-for-legal-practice/", "changes": "Launched at the NBA-SLP Annual General Meeting alongside IT implementation, privacy and cybersecurity guidance." }, { "version": "Circulated text (watermarked DRAFT), 34pp", "date": "2024-11-09T00:00:00.000Z", "url": "https://web.archive.org/web/20241109105437/https://nbaslp.org/wp-content/uploads/2024/04/Guidelines-for-the-Use-of-Artificial-Intelligence-in-the-Nigerian-Legal-Profession.pdf", "changes": "Archived copy of the PDF formerly hosted at nbaslp.org; the live URL no longer resolves after a site rebuild." } ], "summary": "Nigeria's only profession-specific AI guidance. Issued by the NBA Section on Legal Practice on 28 August 2024, it treats AI as augmentative, insists on human oversight, and sets expectations on confidentiality, competence, critical review of outputs, client transparency and consent, supervision of staff, vendor selection and NDPA 2023 compliance. Non-binding.", "keyProvisions": [ "Section 7.1: outputs generated by AI tools must be thoroughly reviewed before reliance in client work, verifying factual accuracy, legal soundness and potential biases.", "Section 7.1: client information, including queries and uploaded documents, should never be used with generative-AI tools lacking robust security and confidentiality protocols.", "Section 7.1: lawyers should disclose use of generative-AI tools to clients and obtain informed consent before proceeding.", "Section 7.1: lawyers remain responsible for subordinates' work and should supervise junior lawyers and staff using generative AI.", "Section 7.2: prohibits overreliance and uninformed use — AI must not replace professional judgment, and lawyers remain ultimately responsible for client matters.", "Sections 6.1 and 10.2: criteria for selecting AI tools and contractual agreements with developers, including data-protection agreements; Appendix B provides an AI impact assessment checklist." ], "relatedIncidents": [], "sources": [ { "label": "NBA launch announcement (official)", "url": "https://blog.nigerianbar.org.ng/2024/09/14/nba-launches-new-technology-guidelines-for-legal-practice/" }, { "label": "Guidelines PDF (official file, archived)", "url": "https://web.archive.org/web/20241109105437/https://nbaslp.org/wp-content/uploads/2024/04/Guidelines-for-the-Use-of-Artificial-Intelligence-in-the-Nigerian-Legal-Profession.pdf" } ], "country": "NG", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision", "client-consent-transparency", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "co-sic-circular-externa-002-2024", "url": "https://safelegalai.com/regulation/other/co-sic-circular-externa-002-2024", "title": "Circular Externa 002 de 2024: Guidelines on the processing of personal data in AI systems", "body": "Superintendence of Industry and Commerce (Superintendencia de Industria y Comercio)", "bodyShort": "SIC", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "risk-classification", "record-keeping" ], "effectiveDate": "2024-08-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-08-21T00:00:00.000Z", "url": "https://sedeelectronica.sic.gov.co/sites/default/files/normativa/Circular%20Externa%20No.%20002%20del%2021%20de%20agosto%20de%202024.pdf" } ], "summary": "Binding instruction from Colombia's data-protection authority to all controllers it supervises: personal-data processing in AI systems must satisfy suitability, necessity, reasonableness and strict proportionality under Laws 1266/2008 and 1581/2012, with prior risk identification and mitigation, privacy by design and by default, staff training and continuous monitoring of security measures.", "keyProvisions": [ "Four criteria: suitability, necessity, reasonableness and strict proportionality.", "Risk identification and mitigation before deploying AI systems.", "Privacy by design and by default, with transparency to data subjects.", "Continuous staff training and monitoring of security measures." ], "relatedIncidents": [], "sources": [ { "label": "SIC Sede Electronica — normative record", "url": "https://sedeelectronica.sic.gov.co/transparencia/normativa/circular-externa-2-de-2024-de-la-superintendencia-de-industria-y-comercio-lineamientos-sobre-el-tratamiento-de-datos" }, { "label": "Circular Externa 002/2024 (PDF)", "url": "https://sedeelectronica.sic.gov.co/sites/default/files/normativa/Circular%20Externa%20No.%20002%20del%2021%20de%20agosto%20de%202024.pdf" } ], "country": "CO", "categories": [ "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "nyc-bar-formal-opinion-2024-5", "url": "https://safelegalai.com/regulation/us-state/nyc-bar-formal-opinion-2024-5", "title": "New York City Bar Formal Opinion 2024-5: Generative Artificial Intelligence", "body": "New York City Bar Association Committee on Professional Ethics", "bodyShort": "NYC Bar", "jurisdiction": "us-state", "region": "New York", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "supervision", "competence" ], "effectiveDate": "2024-08-07T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-08-07T00:00:00.000Z", "url": "https://www.nycbar.org/reports/formal-opinion-2024-5-generative-artificial-intelligence/" } ], "summary": "Formal Opinion 2024-5 from the New York City Bar Association, issued 7 August 2024, applies the New York Rules of Professional Conduct to lawyers' use of generative AI. Lawyers must understand the tools they use, verify all output, protect confidential client information from disclosure to public tools, supervise staff use, and consider whether client disclosure is warranted.", "keyProvisions": [ "Competence requires understanding how a generative AI tool works and its limitations.", "Output must be independently verified before it is relied on or filed.", "Confidential information should not be entered into tools that may retain or train on it.", "Supervisory lawyers must set policies for AI use by lawyers and nonlawyer staff." ], "relatedIncidents": [], "sources": [ { "label": "New York City Bar — Formal Opinion 2024-5", "url": "https://www.nycbar.org/reports/formal-opinion-2024-5-generative-artificial-intelligence/" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "co-corte-constitucional-t-323-2024", "url": "https://safelegalai.com/regulation/other/co-corte-constitucional-t-323-2024", "title": "Sentencia T-323 de 2024: due process and judges' use of generative AI (ChatGPT)", "body": "Constitutional Court of Colombia (Corte Constitucional)", "bodyShort": "Corte Constitucional", "jurisdiction": "other", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "supervision", "disclosure" ], "effectiveDate": "2024-08-02T00:00:00.000Z", "versions": [ { "version": "Sala Segunda de Revision, MP Juan Carlos Cortes Gonzalez", "date": "2024-08-02T00:00:00.000Z", "url": "https://www.corteconstitucional.gov.co/relatoria/2024/T-323-24.htm" } ], "summary": "Reviewing a tutela in which a Cartagena circuit judge used ChatGPT to help reason a health-rights ruling for a child with autism, the Court found no breach of due process on the facts but set criteria for judicial AI use and ordered the Superior Council of the Judiciary to issue guidelines within four months.", "keyProvisions": [ "AI may not replace the judge's reasoning; the natural-judge guarantee requires a human decision.", "Criteria: transparency, accountability, privacy, verification, risk prevention and non-discrimination.", "Judges should disclose and verify any AI use in a decision.", "Order to the Consejo Superior de la Judicatura to issue guidelines, which produced Acuerdo PCSJA24-12243." ], "relatedIncidents": [], "sources": [ { "label": "Corte Constitucional relatoria", "url": "https://www.corteconstitucional.gov.co/relatoria/2024/T-323-24.htm" }, { "label": "Summary (Universidad Externado)", "url": "https://propintel.uexternado.edu.co/resumen-de-la-sentencia-t-323-de-2024-de-la-corte-constitucional-de-colombia-sobre-el-uso-de-ia-por-jueces-de-la-republica/" } ], "country": "CO", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ng-national-ai-strategy-2024", "url": "https://safelegalai.com/regulation/other/ng-national-ai-strategy-2024", "title": "National Artificial Intelligence Strategy (NAIS), August 2024", "body": "Federal Ministry of Communications, Innovation and Digital Economy, with NITDA and the National Centre for Artificial Intelligence and Robotics", "bodyShort": "FMCIDE / NITDA-NCAIR", "jurisdiction": "other", "type": "policy", "status": "proposed", "appliesTo": [ "all" ], "requirements": [ "permissive" ], "versions": [ { "version": "Draft, August 2024", "date": "2024-08-01T00:00:00.000Z", "url": "https://ncair.nitda.gov.ng/wp-content/uploads/2024/08/National-AI-Strategy_01082024-copy.pdf", "changes": "Published for stakeholder input; the hosted PDF remains watermarked DRAFT as at 4 September 2026. Month-only date: the source gives the month, so the 1st is recorded as the date." } ], "summary": "Nigeria's national AI policy framework, published August 2024 and still marked draft. It sets principles of responsible conduct, inclusivity, innovation, sustainability, collaboration and data ethics across five pillars covering infrastructure, ecosystem, adoption, ethical development and governance. It does not mention the judiciary, the legal profession or the administration of justice.", "keyProvisions": [ "Five pillars: AI infrastructure, ecosystem, sectoral adoption, ethical and responsible development, and governance.", "Prepared by FMCIDE with NITDA, NCAIR, NCC, Galaxy Backbone and Data Science Nigeria.", "Contains no reference to courts, the judiciary, legal services or the administration of justice." ], "relatedIncidents": [], "sources": [ { "label": "National AI Strategy August 2024 (NCAIR/NITDA official PDF)", "url": "https://ncair.nitda.gov.ng/wp-content/uploads/2024/08/National-AI-Strategy_01082024-copy.pdf" }, { "label": "FMCIDE NAIS initiative page", "url": "https://fmcide.gov.ng/initiative/nais/" } ], "country": "NG", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ae-uae-charter-development-use-ai", "url": "https://safelegalai.com/regulation/gulf/ae-uae-charter-development-use-ai", "title": "The UAE Charter for the Development and Use of Artificial Intelligence", "body": "UAE Office for Artificial Intelligence, Digital Economy and Remote Work Applications", "bodyShort": "UAE AI Office", "jurisdiction": "gulf", "type": "policy", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "permissive", "supervision", "confidentiality" ], "versions": [ { "version": "Original", "date": "2024-07-30T00:00:00.000Z", "url": "https://u.ae/en/about-the-uae/strategies-initiatives-and-awards/policies/Ai/The-UAE-Charter-for-the-Development-and-Use-of-Artificial-Intelligence", "changes": "Published by the UAE AI Office." } ], "summary": "A non-binding national charter published on 30 July 2024 setting twelve ethical principles for developing and using AI in the UAE, including safety, bias mitigation, data privacy, transparency, human oversight, governance and accountability, and compliance with local law and international treaties. It is the UAE's principal horizontal AI instrument in the absence of an AI statute.", "keyProvisions": [ "Twelve principles covering human-machine relationships, safety, algorithmic bias, data privacy, transparency, human oversight, governance and accountability.", "Requires human judgment to remain over AI systems.", "Expressly non-binding; supplements rather than replaces sectoral law.", "Applies to federal government AI use, which includes the Ministry of Justice's judicial AI services." ], "relatedIncidents": [], "sources": [ { "label": "UAE Government — The UAE Charter for the Development and Use of Artificial Intelligence (official)", "url": "https://u.ae/en/about-the-uae/strategies-initiatives-and-awards/policies/Ai/The-UAE-Charter-for-the-Development-and-Use-of-Artificial-Intelligence" } ], "country": "AE", "categories": [ "ai-act-horizontal", "court-own-use", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "aba-formal-opinion-512", "url": "https://safelegalai.com/regulation/us-federal/aba-formal-opinion-512", "title": "ABA Formal Opinion 512: Generative Artificial Intelligence Tools", "body": "American Bar Association Standing Committee on Ethics and Professional Responsibility", "bodyShort": "ABA", "jurisdiction": "us-federal", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "consent", "verification", "supervision" ], "effectiveDate": "2024-07-29T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-07-29T00:00:00.000Z", "url": "https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf" } ], "summary": "The American Bar Association's first formal ethics opinion on generative AI, issued 29 July 2024, applies the Model Rules to lawyers across the United States. It covers competence, confidentiality with informed client consent for self-learning tools, candor to tribunals, supervision of staff and vendors, and fees. It is permissive with cautions and has no successor opinion.", "keyProvisions": [ "Lawyers must understand the benefits and risks of the generative AI tools they use (Rule 1.1).", "Informed client consent is required before entering confidential information into a self-learning tool (Rule 1.6).", "Output must be verified before it is relied on or filed (Rules 1.1, 3.3).", "Managerial and supervisory lawyers must establish policies for AI use by staff (Rules 5.1, 5.3).", "Lawyers may not bill clients for time saved by AI or for learning to use a tool (Rule 1.5)." ], "relatedIncidents": [ "mata-v-avianca", "park-v-kim" ], "sources": [ { "label": "ABA — Formal Opinion 512 (PDF)", "url": "https://www.americanbar.org/content/dam/aba/administrative/professional_responsibility/ethics-opinions/aba-formal-opinion-512.pdf" } ], "country": "US", "categories": [ "bar-guidance", "competence-duty", "confidentiality-client-data", "client-consent-transparency", "verification-duty", "supervision", "fees-billing", "procurement-vendor" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "au-qld-sc-pd-14-2024-expert-evidence-genai", "url": "https://safelegalai.com/regulation/australia/au-qld-sc-pd-14-2024-expert-evidence-genai", "title": "Practice Direction Number 14 of 2024: Expert evidence in criminal proceedings (other than sentences)", "body": "Supreme Court of Queensland", "bodyShort": "Supreme Court of Queensland", "jurisdiction": "australia", "region": "Queensland", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "parties", "providers" ], "requirements": [ "disclosure", "record-keeping", "verification" ], "effectiveDate": "2024-07-15T00:00:00.000Z", "versions": [ { "version": "Commenced 15 July 2024 (subsequently amended)", "date": "2024-07-15T00:00:00.000Z", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/797050/scpd-14-of-2024.pdf" } ], "summary": "Queensland's expert-evidence practice direction for criminal proceedings contains the most granular AI disclosure requirement in Australia: where an expert used generative AI to formulate or express an opinion, the report must name the program used and annex complete records of the prompts given and of the outputs delivered.", "keyProvisions": [ "Where generative AI assisted in formulating or expressing an opinion, the report must specify the name of the program used and how it was used.", "The report must annex a complete record of the inputs (prompts), including source material, default values and variable sets.", "The report must annex a complete record of the outputs delivered by the program.", "Applies to all criminal proceedings commenced by indictment presented on or after 15 July 2024, other than sentences." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Queensland — Practice Direction 14 of 2024 (official PDF)", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/797050/scpd-14-of-2024.pdf" } ], "country": "AU", "categories": [ "evidence-admissibility", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hk-judiciary-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/hong-kong/hk-judiciary-generative-ai-guidelines", "title": "Guidelines on the Use of Generative AI for Judges, Judicial Officers and Support Staff", "body": "Hong Kong Judiciary", "bodyShort": "Hong Kong Judiciary", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality" ], "effectiveDate": "2024-07-01T00:00:00.000Z", "versions": [ { "version": "July 2024", "date": "2024-07-01T00:00:00.000Z", "url": "https://www.judiciary.hk/doc/en/court_services_facilities/guidelines_on_the_use_of_generative_ai.pdf" } ], "summary": "Guidelines issued by the Hong Kong Judiciary in July 2024 for judges, judicial officers and support staff on generative AI. Using generative AI for legal analysis or reasoning is not recommended, administrative uses are permitted, confidential case information must not be entered into public tools, and any output must be verified. Sectoral guidelines for the legal profession are under consultation.", "keyProvisions": [ "Generative AI for legal analysis, research or reasoning in judicial work is not recommended.", "Administrative and drafting assistance uses are permitted with human review.", "Confidential or sensitive case material must not be entered into public tools.", "Judicial officers remain responsible for all output used." ], "relatedIncidents": [], "article": "tier3-ai-guidance-india-brazil-hongkong-gulf", "sources": [ { "label": "Hong Kong Judiciary — generative AI guidelines (PDF)", "url": "https://www.judiciary.hk/doc/en/court_services_facilities/guidelines_on_the_use_of_generative_ai.pdf" } ], "country": "HK", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "court-own-use", "disclosure-filings" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-mn-msba-ai-working-group-report", "url": "https://safelegalai.com/regulation/us-state/us-mn-msba-ai-working-group-report", "title": "Minnesota State Bar Association AI Working Group: Final Report and Recommendations", "body": "Minnesota State Bar Association, Artificial Intelligence Working Group", "bodyShort": "Minnesota State Bar Association", "jurisdiction": "us-state", "region": "Minnesota", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms", "providers" ], "requirements": [ "competence", "confidentiality", "verification" ], "versions": [ { "version": "Final Report and Recommendations", "date": "2024-07-01T00:00:00.000Z", "url": "https://mnbars.org/?pg=MSBAAnnouncements&blAction=showEntry&blogEntry=112704", "changes": "Adopted June/July 2024; the source gives a month only (adopted June/July 2024), so the day is recorded as the 1st" } ], "summary": "Minnesota's bar produced a working-group report rather than an ethics opinion. It recommends an AI regulatory sandbox for access-to-justice LLM tools and sets out guidance on lawyer use of generative AI under the Minnesota Rules of Professional Conduct. Text not verified in this run.", "keyProvisions": [ "Recommends an AI regulatory sandbox for access-to-justice large language model tools", "Provides guidance on lawyer use of generative AI under the Minnesota Rules of Professional Conduct", "Bar association report, not an ethics opinion; the LPRB has issued no numbered AI opinion" ], "relatedIncidents": [], "sources": [ { "label": "MSBA AI Working Group Report (official announcement)", "url": "https://mnbars.org/?pg=MSBAAnnouncements&blAction=showEntry&blogEntry=112704" } ], "country": "US", "categories": [ "bar-guidance", "competence-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-mo-clay-county-local-rule-3-3-1", "url": "https://safelegalai.com/regulation/us-state/us-mo-clay-county-local-rule-3-3-1", "title": "Missouri 7th Judicial Circuit (Clay County) Local Rule 3.3.1: Disclosure of Artificial Intelligence Use", "body": "Seventh Judicial Circuit of Missouri (Clay County)", "bodyShort": "MO 7th Judicial Circuit", "jurisdiction": "us-state", "region": "Missouri", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "certification" ], "versions": [ { "version": "Local Rule 3.3.1", "date": "2024-07-01T00:00:00.000Z", "url": "http://www.circuit7.net/circuitcourt/Local-Court-Rules.pdf", "changes": "Adopted July 2024; the source gives a month only (adopted July 2024), so the day is recorded as the 1st" } ], "summary": "Missouri has no statewide AI filing rule, so several circuits have made their own. Clay County's Local Rule 3.3.1 requires disclosure of artificial intelligence use in filings from July 2024. Its operative wording was not read in this research run and must be confirmed.", "keyProvisions": [ "Requires disclosure of artificial intelligence use in filings in the 7th Judicial Circuit (terms not yet verified)", "One of at least four Missouri circuits with local AI disclosure rules", "No statewide Missouri Supreme Court AI filing rule exists" ], "relatedIncidents": [], "sources": [ { "label": "7th Judicial Circuit local court rules (PDF)", "url": "http://www.circuit7.net/circuitcourt/Local-Court-Rules.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "tr-ai-bill-2024", "url": "https://safelegalai.com/regulation/other/tr-ai-bill-2024", "title": "Artificial Intelligence Bill [Yapay Zekâ Kanunu Teklifi]", "body": "Grand National Assembly of Türkiye [Türkiye Büyük Millet Meclisi]", "bodyShort": "TBMM", "jurisdiction": "other", "type": "statute", "status": "proposed", "appliesTo": [ "providers", "all" ], "requirements": [ "risk-classification" ], "versions": [ { "version": "As submitted, 28th legislative term", "date": "2024-06-25T00:00:00.000Z", "url": "https://www.tbmm.gov.tr/Yasama/KanunTeklifi/e21539a0-888a-4500-81be-01904a918c53", "changes": "Referred on submission to the Industry, Trade, Energy, Natural Resources, Information and Technology Committee as lead committee and the Justice Committee as secondary committee. Status recorded as in committee as at 4 Sep 2026." } ], "summary": "Türkiye's first general AI bill, submitted to the Grand National Assembly on 25 June 2024 with Ömer Faruk Gergerlioğlu as first signatory. It has not been reported out of committee. It contains no justice-sector or legal-practice risk classification, and no horizontal AI statute is in force in Türkiye as at 4 September 2026.", "keyProvisions": [ "Submitted 25 June 2024; lead committee is the Industry, Trade, Energy, Natural Resources, Information and Technology Committee.", "The Justice Committee is named as a secondary (tali) committee, but the bill does not classify judicial or legal-services uses.", "Still in committee with no committee decision date recorded as at 4 September 2026." ], "relatedIncidents": [], "sources": [ { "label": "TBMM bill record", "url": "https://www.tbmm.gov.tr/Yasama/KanunTeklifi/e21539a0-888a-4500-81be-01904a918c53" } ], "country": "TR", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ch-sav-wegleitung-ki", "url": "https://safelegalai.com/regulation/other/ch-sav-wegleitung-ki", "title": "SAV-Wegleitung für den Umgang mit künstlicher Intelligenz", "body": "Schweizerischer Anwaltsverband / Fédération Suisse des Avocats (Swiss Bar Association)", "bodyShort": "SAV/FSA", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "consent", "disclosure", "permissive" ], "versions": [ { "version": "Approved by the SAV board", "date": "2024-06-14T00:00:00.000Z", "url": "https://www.sav-fsa.ch/documents/672183/2025869/SAV_KI+Guidelines+2024.pdf/68d5eab4-c2ad-50bd-506f-7ba6af15a839", "changes": "Original guideline approved by the SAV Vorstand on 14 June 2024." }, { "version": "Updated version", "date": "2025-02-16T00:00:00.000Z", "url": "https://digital.sav-fsa.ch/documents/1060627/0/SAV-Wegleitung_f%C3%BCr_den_Umgang_mit_k%C3%BCnstlicher_Intelligenz_16.2.25.pdf/0b498c8a-85aa-e089-32a2-18da59ee658c", "changes": "Updated to reflect the Federal Council's 12 February 2025 decisions on AI regulation and the BAKOM Auslegeordnung." } ], "summary": "The Swiss Bar Association's guideline on AI, approved 14 June 2024 and updated 16 February 2025. It focuses on generative AI, sets a three-route test for putting client information into a tool, requires independent verification of output, addresses liability, copyright and notification duties, and explains the Swiss, EU and Council of Europe regulatory position.", "keyProvisions": [ "Before using any AI, clarify what happens to input data, who can access it and where it is stored, having regard to the Anwaltsgeheimnis, the DSG and the Standesregeln.", "Only three routes permit confidential input: an on-premise install where no data leaves the firm network; compliance with the SAV's IT-outsourcing and cloud rules; or an informed consent-and-waiver from the client.", "Output must be checked independently and critically; an AI cannot verify its own output. Failure sources named are hallucination, missing or dated training data, sycophancy and bias.", "The lawyer is liable for defective performance and cannot plead that the AI made the mistake; a liability limitation should be agreed with the client where lawful.", "A duty to disclose AI use may arise from the provider's terms, or where the client requires or legitimately expects personal performance.", "Switzerland will not transpose the EU AI Act; the Federal Council decided on 12 February 2025 to adopt the Council of Europe AI Convention and legislate sector-specifically, with an EJPD consultation draft due by the end of 2026." ], "relatedIncidents": [], "sources": [ { "label": "SAV — Wegleitung für den Umgang mit künstlicher Intelligenz, 16 Feb 2025 (official PDF)", "url": "https://digital.sav-fsa.ch/documents/1060627/0/SAV-Wegleitung_f%C3%BCr_den_Umgang_mit_k%C3%BCnstlicher_Intelligenz_16.2.25.pdf/0b498c8a-85aa-e089-32a2-18da59ee658c" }, { "label": "SAV — Wegleitung Künstliche Intelligenz (landing page, official)", "url": "https://www.sav-fsa.ch/testseite/-/asset_publisher/DbkRv9dMWxix/content/wegleitung-k%C3%BCnstliche-intelligenz-ki-" } ], "country": "CH", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "client-consent-transparency", "procurement-vendor", "supervision", "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-jams-ai-disputes-rules", "url": "https://safelegalai.com/regulation/international/int-jams-ai-disputes-rules", "title": "JAMS Rules Governing Disputes Involving Artificial Intelligence Systems", "body": "JAMS", "bodyShort": "JAMS", "jurisdiction": "international", "type": "court-rule", "status": "in-force", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "confidentiality", "record-keeping" ], "effectiveDate": "2024-06-14T00:00:00.000Z", "versions": [ { "version": "Effective", "date": "2024-06-14T00:00:00.000Z", "url": "https://www.jamsadr.com/artificial-intelligence-disputes-clause-and-rules", "changes": "First ADR rules specific to disputes about AI systems; announced 23 April 2024." } ], "summary": "JAMS rules and model clause for arbitrating disputes about AI systems. They are procedural rules for AI subject-matter disputes rather than guidance on participants' use of AI: JAMS proposes panelists with technical expertise, and a default AI Disputes Protective Order limits inspection of AI systems, models and training data to experts in a secured environment.", "keyProvisions": [ "Rule 15(b): JAMS proposes only panelists approved for disputes involving technical subject matter.", "Rule 16.1(a): default AI Disputes Protective Order applies unless the parties agree otherwise.", "Rule 16.1(b): production of AI systems and related materials is limited to experts in a secured environment, with no removal of materials." ], "relatedIncidents": [], "sources": [ { "label": "JAMS — AI Disputes Clause and Rules", "url": "https://www.jamsadr.com/artificial-intelligence-disputes-clause-and-rules" } ], "country": "INT", "categories": [ "arbitration-adr", "evidence-admissibility" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ks-shawnee-county-dcr-3-125", "url": "https://safelegalai.com/regulation/us-state/us-ks-shawnee-county-dcr-3-125", "title": "Shawnee County District Court Rule 3.125: Pleadings Using Generative Artificial Intelligence", "body": "Third Judicial District of Kansas (Shawnee County District Court)", "bodyShort": "Shawnee County District Court", "jurisdiction": "us-state", "region": "Kansas", "type": "court-rule", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "certification", "verification" ], "effectiveDate": "2024-06-14T00:00:00.000Z", "versions": [ { "version": "DCR 3.125", "date": "2024-06-14T00:00:00.000Z", "url": "https://www.shawneecourt.org/428/DCR-3125-Pleadings-Using-Generative-Arti", "changes": "Effective 14 June 2024" } ], "summary": "Shawnee County District Court requires both verification and disclosure. Any filer who used generative AI for any portion of a pleading must check the output for accuracy against print reporters or traditional databases and must state on the first page, in writing, that the document contains AI-generated content, with a certification of the accuracy check. Sanctions may follow a violation.", "keyProvisions": [ "Applies where any portion of a pleading or other paper was drafted using generative AI, naming ChatGPT, Harvey.AI and Google Bard as examples", "Filer must verify that any generated language, materials or content was checked for accuracy using print reporters, traditional legal databases or other reliable means", "Filer must disclose to the court and opposing parties at the time of filing, in writing and clearly marked on the first page, that the document contains AI-generated content", "The disclosure must include a certification that the content was checked for accuracy as required by the rule", "AI systems may not be factually or legally trustworthy sources of information without human verification", "Violation may subject attorneys and self-represented litigants to sanctions" ], "relatedIncidents": [], "sources": [ { "label": "Shawnee County DCR 3.125", "url": "https://www.shawneecourt.org/428/DCR-3125-Pleadings-Using-Generative-Arti" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty", "sanctions-record" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-wv-legal-ethics-opinion-24-01", "url": "https://safelegalai.com/regulation/us-state/us-wv-legal-ethics-opinion-24-01", "title": "Legal Ethics Opinion 24-01: Artificial Intelligence", "body": "West Virginia Lawyer Disciplinary Board", "bodyShort": "WV Lawyer Disciplinary Board", "jurisdiction": "us-state", "region": "West Virginia", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "supervision" ], "versions": [ { "version": "Original", "date": "2024-06-14T00:00:00.000Z", "url": "https://storage.googleapis.com/msgsndr/Rgd68xOkcVdteTsBkf6O/media/667ac9c219bb7a1f7a4df4c2.pdf", "changes": "Initial publication" } ], "summary": "LEI 24-01 holds that nothing in the West Virginia Rules of Professional Conduct per se prohibits AI use in law practice. It distinguishes non-generative from generative AI, maps the main legal use cases, and requires lawyers to exercise independent judgment, communicate with clients, maintain confidentiality, keep fees reasonable, comply with advertising rules and supervise AI-assisted work.", "keyProvisions": [ "Nothing in the Rules of Professional Conduct per se prohibits a lawyer from using AI as a tool in law practice", "Lawyers must check jurisdiction-specific rules and standing orders on whether generative AI is prohibited, whether disclosure is required, and whether human verification must be certified", "Lawyers must exercise independent judgment and not delegate it to an AI tool", "Efficiency gains must not result in double billing or falsely inflated claims of time (Rules 1.5, 8.4(c), 8.4(d))", "Client confidentiality must be maintained, with particular caution about generative AI intake chatbots", "AI-assisted work must be supervised as with any other assistance" ], "relatedIncidents": [], "sources": [ { "label": "West Virginia Legal Ethics Opinion 24-01 (official PDF)", "url": "https://storage.googleapis.com/msgsndr/Rgd68xOkcVdteTsBkf6O/media/667ac9c219bb7a1f7a4df4c2.pdf" } ], "country": "US", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "fees-billing", "competence-duty", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hk-pcpd-ai-model-personal-data-protection-framework", "url": "https://safelegalai.com/regulation/hong-kong/hk-pcpd-ai-model-personal-data-protection-framework", "title": "Artificial Intelligence: Model Personal Data Protection Framework", "body": "Office of the Privacy Commissioner for Personal Data, Hong Kong", "bodyShort": "PCPD", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "risk-classification", "confidentiality", "supervision", "record-keeping" ], "effectiveDate": "2024-06-11T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-06-11T00:00:00.000Z", "url": "https://www.pcpd.org.hk/english/resources_centre/publications/files/ai_protection_framework.pdf" } ], "summary": "Published on 11 June 2024, the PCPD's Model Framework gives organisations that procure, implement and use AI, including generative AI, recommendations for complying with the Personal Data (Privacy) Ordinance. It builds on the PCPD's 2021 guidance on the ethical development and use of AI and is structured in four parts.", "keyProvisions": [ "Part I: establish AI strategy and governance, including governance considerations for procuring AI solutions, an AI governance committee, and employee training.", "Part II: conduct risk assessment and set the level of human oversight using a risk-based approach.", "Part III: prepare and manage data for customisation and use of AI systems, and test, validate and continuously monitor the systems.", "Part IV: communicate and engage with stakeholders, including information provision, data subject rights and explainable AI." ], "relatedIncidents": [], "sources": [ { "label": "PCPD — Model Framework (official PDF)", "url": "https://www.pcpd.org.hk/english/resources_centre/publications/files/ai_protection_framework.pdf" }, { "label": "PCPD — media statement (11 June 2024)", "url": "https://www.pcpd.org.hk/english/news_events/media_statements/press_20240611.html" } ], "country": "HK", "categories": [ "data-protection", "procurement-vendor", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ireland-judicial-council-generative-ai-guidelines-judges", "url": "https://safelegalai.com/regulation/ireland/ireland-judicial-council-generative-ai-guidelines-judges", "title": "Guidelines for Judges: The Use of Generative Artificial Intelligence", "body": "Judicial Council of Ireland", "bodyShort": "Judicial Council of Ireland", "jurisdiction": "ireland", "type": "guidance", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "verification", "confidentiality", "prohibition" ], "effectiveDate": "2024-06-01T00:00:00.000Z", "versions": [ { "version": "June 2024 (published November 2025)", "date": "2024-06-01T00:00:00.000Z", "url": "https://judicialcouncil.ie/assets/uploads/AI%20Guidance%20-%20Guidelines%20For%20Judges%20(June%202024).pdf" } ], "summary": "Guidelines from the Judicial Council of Ireland, dated June 2024 and published in November 2025 as a living document, on Irish judges' use of generative AI. Judges may use AI for routine and administrative tasks only, must verify any output, must not enter confidential case material into public tools, and must not use AI to decide cases or draft reasoning.", "keyProvisions": [ "Generative AI may be used for routine and administrative tasks only.", "Judges must verify any AI output before relying on it.", "Confidential or sensitive case information must not be entered into public tools.", "AI must not be used to determine outcomes or draft judicial reasoning." ], "relatedIncidents": [], "article": "ireland-ai-guidance-history", "sources": [ { "label": "Judicial Council of Ireland — guidelines for judges (PDF)", "url": "https://judicialcouncil.ie/assets/uploads/AI%20Guidance%20-%20Guidelines%20For%20Judges%20(June%202024).pdf" }, { "label": "Judicial Council of Ireland — news announcement", "url": "https://judicialcouncil.ie/news/guidelines-for-judges-the-use-of-generative-artificial-intelligence/" } ], "country": "IE", "categories": [ "judicial-use", "ai-decision-prohibition", "verification-duty", "confidentiality-client-data", "court-own-use" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "us-ca5-proposed-rule-32-3-genai-certification", "url": "https://safelegalai.com/regulation/us-federal/us-ca5-proposed-rule-32-3-genai-certification", "title": "Proposed amendment to Fifth Circuit Rule 32.3 and Form 6 (generative AI certification)", "body": "United States Court of Appeals for the Fifth Circuit", "bodyShort": "Fifth Circuit", "jurisdiction": "us-federal", "type": "court-rule", "status": "withdrawn", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "certification" ], "versions": [ { "version": "Court decision not to adopt", "date": "2024-06-01T00:00:00.000Z", "url": "https://www.ca5.uscourts.gov/docs/default-source/default-document-library/court-decision-on-proposed-rule.pdf" } ], "summary": "In late 2023 the Fifth Circuit proposed requiring every filer to certify either that no generative AI was used or that all AI-generated text had been reviewed by a human. After public comment the court announced in June 2024 that it would not adopt the rule, relying on existing Rule 11 and candor obligations. No AI-specific rule applies.", "keyProvisions": [ "Would have required a certificate in every brief about the use of generative AI.", "Not adopted; the court concluded existing rules already govern accuracy and candor.", "Parties remain bound by Rule 11 and professional conduct rules for any AI-assisted filing." ], "relatedIncidents": [], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "Fifth Circuit — court decision on proposed rule (PDF)", "url": "https://www.ca5.uscourts.gov/docs/default-source/default-document-library/court-decision-on-proposed-rule.pdf" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "qls-guidance-statement-37-ai", "url": "https://safelegalai.com/regulation/australia/qls-guidance-statement-37-ai", "title": "Guidance Statement No. 37: Artificial Intelligence in Legal Practice", "body": "Queensland Law Society", "bodyShort": "QLS", "jurisdiction": "australia", "region": "Queensland", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "competence", "confidentiality", "verification", "supervision" ], "effectiveDate": "2024-05-31T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-05-31T00:00:00.000Z", "url": "https://www.qls.com.au/practising-law-in-qld/ethics/guidance-statements/no-37-artificial-intelligence-in-legal-practice", "archiveUrl": "https://web.archive.org/web/20260904134130/https://www.qls.com.au/practising-law-in-qld/ethics/guidance-statements/no-37-artificial-intelligence-in-legal-practice" } ], "summary": "Guidance Statement No. 37 from the Queensland Law Society, issued 31 May 2024, applies the Australian Solicitors Conduct Rules to solicitors' use of AI. It is permissive with cautions: solicitors must be competent in the tools they use, protect confidentiality, verify output before reliance, supervise staff use, and consider whether to inform clients about AI use.", "keyProvisions": [ "Solicitors must understand the AI tools they use and their limitations.", "Confidential client information must be protected when using AI services.", "AI-generated content must be verified before it is relied on or filed.", "Principals must supervise AI use by employees and set firm policies." ], "relatedIncidents": [], "article": "australia-law-society-ai-guidance-history", "sources": [ { "label": "Queensland Law Society — Guidance Statement No. 37", "url": "https://www.qls.com.au/practising-law-in-qld/ethics/guidance-statements/no-37-artificial-intelligence-in-legal-practice" } ], "country": "AU", "categories": [ "bar-guidance", "competence-duty", "confidentiality-client-data", "verification-duty", "supervision", "procurement-vendor", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-pa-joint-formal-opinion-2024-200", "url": "https://safelegalai.com/regulation/us-state/us-pa-joint-formal-opinion-2024-200", "title": "Joint Formal Opinion 2024-200: Ethical Issues Regarding the Use of Artificial Intelligence", "body": "Pennsylvania Bar Association Committee on Legal Ethics and Professional Responsibility and Philadelphia Bar Association Professional Guidance Committee", "bodyShort": "PBA and Philadelphia Bar", "jurisdiction": "us-state", "region": "Pennsylvania", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "consent", "supervision" ], "versions": [ { "version": "Original", "date": "2024-05-22T00:00:00.000Z", "url": "https://www.pabar.org/Members/catalogs/Ethics%20Opinions/Formal/Joint%20Formal%20Opinion%202024-200.pdf", "changes": "Initial publication" } ], "summary": "A joint opinion of the Pennsylvania and Philadelphia bar ethics committees concluding that, with appropriate safeguards, lawyers may use AI consistently with the Pennsylvania Rules of Professional Conduct. It sets out twelve best practices spanning truthfulness, citation verification, competence, confidentiality, conflicts, client communication, bias, misuse, billing, professional judgment and transparency.", "keyProvisions": [ "With appropriate safeguards, lawyers may use AI in accordance with the Pennsylvania Rules of Professional Conduct", "Lawyers must be competent in the use of AI technologies and must verify that citations accurately reflect the content referenced", "Lawyers must safeguard client information and ensure AI systems handling it adhere to strict confidentiality measures", "Lawyers must communicate with clients about their AI use and, if necessary, obtain consent before using certain AI tools", "AI-related expenses must be reasonable and appropriately disclosed to clients", "Lawyers should be transparent with clients, colleagues and the courts about AI use, including its limitations and uncertainties", "AI does not replace personal review of cases, statutes and other legislative materials" ], "relatedIncidents": [], "sources": [ { "label": "Joint Formal Opinion 2024-200 (official PDF)", "url": "https://www.pabar.org/Members/catalogs/Ethics%20Opinions/Formal/Joint%20Formal%20Opinion%202024-200.pdf" } ], "country": "US", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "fees-billing", "competence-duty", "client-consent-transparency" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-nh-bar-ethics-corner-ai-in-practice", "url": "https://safelegalai.com/regulation/us-state/us-nh-bar-ethics-corner-ai-in-practice", "title": "Ethics of Using Artificial Intelligence in Practice (NHBA Ethics Corner)", "body": "New Hampshire Bar Association Ethics Committee", "bodyShort": "NH Bar Association", "jurisdiction": "us-state", "region": "New Hampshire", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "competence", "confidentiality", "verification" ], "versions": [ { "version": "Ethics Corner article", "date": "2024-05-15T00:00:00.000Z", "url": "https://www.nhbar.org/using-artificial-intelligence-in-practice/", "changes": "Initial publication; companion article on drafting documents with AI" } ], "summary": "A practical ethics article from the New Hampshire Bar Association Ethics Committee applying NH RPC 1.1, 1.6, 3.3, 3.4 and 4.1 to generative AI. It stresses technology competence, review of vendor privacy policies before entering client data, and verification of AI output, citing Mata v. Avianca. It is guidance rather than a formal opinion.", "keyProvisions": [ "NH RPC 1.1 requires understanding how an AI system is designed and operates before relying on output", "NH RPC 1.6 requires reviewing vendor terms and privacy policies before inputting client information", "Lawyers must verify the accuracy of AI results and be able to explain them to a court", "Unverified AI content risks Rule 11 sanctions and violations of RPC 3.3, 3.4 and 4.1", "Generative AI should generally not be used to draft pleadings because that usually discloses confidential client information to the tool" ], "relatedIncidents": [], "sources": [ { "label": "NHBA, Ethics of Using Artificial Intelligence in Practice", "url": "https://www.nhbar.org/using-artificial-intelligence-in-practice/" }, { "label": "NHBA, Ethics of Drafting Documents with Artificial Intelligence", "url": "https://www.nhbar.org/drafting-documents-with-artificial-intelligence/" } ], "country": "US", "categories": [ "bar-guidance", "confidentiality-client-data", "verification-duty", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "canada-federal-court-ai-notice", "url": "https://safelegalai.com/regulation/canada/canada-federal-court-ai-notice", "title": "Notice to the Parties and the Profession: The Use of Artificial Intelligence in Court Proceedings", "body": "Federal Court of Canada", "bodyShort": "Federal Court of Canada", "jurisdiction": "canada", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "verification" ], "effectiveDate": "2023-12-20T00:00:00.000Z", "versions": [ { "version": "Updated notice", "date": "2024-05-07T00:00:00.000Z", "url": "https://www.fct-cf.ca/Content/assets/pdf/base/FC-Updated-AI-Notice-EN.pdf", "changes": "Clarifies the scope of the declaration requirement and the meaning of AI-generated content; retains the human-in-the-loop principle." }, { "version": "Original", "date": "2023-12-20T00:00:00.000Z", "url": "https://www.fct-cf.ca/Content/assets/pdf/base/2023-12-20-notice-use-of-ai-in-court-proceedings.pdf" } ], "summary": "The Federal Court of Canada's notice, first issued 20 December 2023 and updated 7 May 2024, requires parties and counsel to declare in the first paragraph of any document containing AI-generated content that AI was used to prepare it, and to verify that content. It was among the first mandatory disclosure rules for court filings in a common-law jurisdiction.", "keyProvisions": [ "A declaration must appear in the first paragraph of any document with AI-generated content submitted to the court.", "Human-in-the-loop verification of AI-generated content is expected.", "Applies to counsel and self-represented litigants.", "The court itself will not use AI in decision-making without prior consultation." ], "relatedIncidents": [ "hussein-v-canada" ], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Federal Court of Canada — updated AI notice (PDF)", "url": "https://www.fct-cf.ca/Content/assets/pdf/base/FC-Updated-AI-Notice-EN.pdf" }, { "label": "Federal Court of Canada — original notice, 20 December 2023 (PDF)", "url": "https://www.fct-cf.ca/Content/assets/pdf/base/2023-12-20-notice-use-of-ai-in-court-proceedings.pdf" } ], "country": "CA", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "israel-bar-ethics-opinion-at-60-24", "url": "https://safelegalai.com/regulation/israel/israel-bar-ethics-opinion-at-60-24", "title": "Preliminary Opinion on the Use of Artificial Intelligence (AI) in Lawyers' Work, decision At/60/24", "body": "Israel Bar Association [Lishkat Orkhei HaDin], National Ethics Committee", "bodyShort": "Israel Bar National Ethics Committee", "jurisdiction": "israel", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "consent", "competence", "disclosure" ], "versions": [ { "version": "Decision At/60/24, preliminary opinion", "date": "2024-05-07T00:00:00.000Z", "url": "https://www.israelbar.biz/halishka_vehmechozot/lishkat_orche_din/mosdot_halishka/hamohza_harzit/545.aspx", "changes": "First Israeli bar instrument on generative AI. Reported by law.co.il on 8 May 2024; some later reports date the circulated text to August 2024. The opinion text was not read in full; the record rests on the Bar's page and secondary analyses, so it is unverified." } ], "summary": "The Israel Bar's first ethics opinion on generative AI, adopted by the National Ethics Committee as decision At/60/24 in 2024. It permits AI use but requires individual verification of AI output against external authority, bars entering client, confidential, privileged or copyright-protected material into AI platforms, and requires written client consent before client personal data enters open systems.", "keyProvisions": [ "Lawyers must individually check the reliability of AI output and corroborate it with external sources; unverifiable material must be flagged to the court.", "Client personal, confidential, privileged and copyright-protected information must not be entered into AI platforms.", "Explicit written client consent is required before client personal information is used in open AI systems, specifying document types and security measures.", "Closed, in-house AI systems are preferred for confidential work to limit unauthorised access.", "The lawyer bears full civil and ethical responsibility for AI use; unverified reliance breaches the duties of loyalty and diligence, and sanctions can run to licence revocation." ], "relatedIncidents": [], "sources": [ { "label": "Israel Bar Association, decision At/60/24 (official page)", "url": "https://www.israelbar.biz/halishka_vehmechozot/lishkat_orche_din/mosdot_halishka/hamohza_harzit/545.aspx" }, { "label": "Opinion PDF hosted on gov.il", "url": "https://www.gov.il/BlobFolder/generalpage/ai_info_law/he/%D7%92%D7%99%D7%9C%D7%95%D7%99%20%D7%93%D7%A2%D7%AA%20%D7%91%D7%A2%D7%A0%D7%99%D7%99%D7%9F%20%D7%A9%D7%99%D7%9E%D7%95%D7%A9%20%D7%91%D7%91%D7%99%D7%A0%D7%94%20%D7%9E%D7%9C%D7%90%D7%9B%D7%95%D7%AA%D7%99%D7%AA%20%D7%91%D7%A2%D7%91%D7%95%D7%93%D7%AA%20%D7%A2%D7%95%D7%A8%D7%9B%D7%99%20%D7%94%D7%93%D7%99%D7%9F%20(2024).pdf" }, { "label": "law.co.il report on publication (8 May 2024)", "url": "https://www.law.co.il/news/2024/05/08/ai-ethcis-rules-for-israeli-lawyers/" }, { "label": "Pearl Cohen analysis", "url": "https://www.pearlcohen.com/israeli-bar-associations-ethical-guidelines-on-using-ai/" } ], "country": "IL", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "client-consent-transparency", "competence-duty", "disclosure-filings", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "de-dsk-orientierungshilfe-ki-datenschutz", "url": "https://safelegalai.com/regulation/other/de-dsk-orientierungshilfe-ki-datenschutz", "title": "Orientierungshilfe: Künstliche Intelligenz und Datenschutz", "body": "Konferenz der unabhängigen Datenschutzaufsichtsbehörden des Bundes und der Länder (German Data Protection Conference)", "bodyShort": "DSK", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "firms" ], "requirements": [ "confidentiality", "consent", "record-keeping", "supervision" ], "versions": [ { "version": "Version 1.0", "date": "2024-05-06T00:00:00.000Z", "url": "https://www.datenschutzkonferenz-online.de/media/oh/20240506_DSK_Orientierungshilfe_KI_und_Datenschutz.pdf", "changes": "First version; the DSK states it will be updated as developments require." } ], "summary": "Guidance of 6 May 2024 from the conference of Germany's federal and state data protection authorities on using AI applications, in particular large language models, in compliance with data protection law. It is addressed primarily to controllers deploying AI and is the reference text the BRAK directs law firms to.", "keyProvisions": [ "Define the field of use and purpose, check its lawfulness, and prefer uses that need no personal data.", "No automated final decision: a human must take the decision (reflecting Art. 22 GDPR).", "Distinguish closed from open systems and be transparent about training use and prompt history, with a choice where possible.", "Carry out a data protection impact assessment, set internal rules, define and bind responsibility, and involve the DPO and staff representatives.", "Protect employees and set up business accounts rather than personal ones; apply privacy by design and by default." ], "relatedIncidents": [], "sources": [ { "label": "DSK — Orientierungshilfe KI und Datenschutz (official PDF)", "url": "https://www.datenschutzkonferenz-online.de/media/oh/20240506_DSK_Orientierungshilfe_KI_und_Datenschutz.pdf" } ], "country": "DE", "categories": [ "data-protection", "ai-decision-prohibition", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "int-oecd-recommendation-ai", "url": "https://safelegalai.com/regulation/international/int-oecd-recommendation-ai", "title": "OECD Recommendation of the Council on Artificial Intelligence (OECD/LEGAL/0449)", "body": "Organisation for Economic Co-operation and Development", "bodyShort": "OECD", "jurisdiction": "international", "type": "policy", "status": "in-force", "appliesTo": [ "all" ], "requirements": [ "permissive" ], "versions": [ { "version": "Revised", "date": "2024-05-03T00:00:00.000Z", "url": "https://legalinstruments.oecd.org/api/print?ids=648&lang=en", "changes": "Updated for generative AI; 47 adherents." }, { "version": "Revised definition of AI system", "date": "2023-11-08T00:00:00.000Z", "url": "https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449", "changes": "Definition later adopted by the EU AI Act and the Ciarb arbitration guideline." }, { "version": "Original", "date": "2019-05-22T00:00:00.000Z", "url": "https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449" } ], "summary": "Non-binding OECD standard adopted in 2019 and revised on 3 May 2024, setting five values-based principles: inclusive growth; respect for the rule of law, human rights and democratic values including fairness and privacy; transparency and explainability; robustness, security and safety; and accountability. Its AI-system definition is used by the EU AI Act.", "keyProvisions": [ "Principle 1.2: AI actors should respect the rule of law, human rights and democratic values throughout the AI lifecycle.", "Principle 1.3: transparency and responsible disclosure so people can understand and challenge AI outcomes.", "Principle 1.5: accountability of AI actors, including traceability and risk management." ], "relatedIncidents": [], "sources": [ { "label": "OECD Legal Instruments — OECD/LEGAL/0449", "url": "https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449" } ], "country": "INT", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "au-qld-courts-judicial-officers-genai-guidelines", "url": "https://safelegalai.com/regulation/australia/au-qld-courts-judicial-officers-genai-guidelines", "title": "The use of generative AI: guidelines for judicial officers", "body": "Queensland Courts", "bodyShort": "Queensland Courts", "jurisdiction": "australia", "region": "Queensland", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "verification", "confidentiality" ], "effectiveDate": "2024-05-01T00:00:00.000Z", "versions": [ { "version": "Companion to the May 2024 non-lawyer guidelines", "date": "2024-05-01T00:00:00.000Z", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0009/879714/the-use-of-generative-ai-guidelines-for-judicial-officers.pdf", "changes": "Only the month is known: the date is inferred from the companion May 2024 guidelines for non-lawyers and recorded as the first of that month. The PDF sits behind a Cloudflare block and was not opened in this research pass." } ], "summary": "Queensland Courts' companion guidelines for judicial officers, published alongside the guidelines for responsible use of generative AI by non-lawyers and indexed on the Law Society of New South Wales national court-protocols tracker. The document could not be opened in this research pass, so its provisions are recorded from the index entry rather than from the text.", "keyProvisions": [ "Companion document to the Queensland Courts guidelines for responsible use of generative AI by non-lawyers.", "Addressed to judicial officers rather than to parties or practitioners." ], "relatedIncidents": [], "sources": [ { "label": "Queensland Courts — guidelines for judicial officers (PDF; blocked at time of access)", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0009/879714/the-use-of-generative-ai-guidelines-for-judicial-officers.pdf" }, { "label": "Law Society of NSW — court protocols on AI (index listing this document)", "url": "https://www.lawsociety.com.au/AI-hub/court-protocols-ai" } ], "country": "AU", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "qld-courts-generative-ai-guidelines-non-lawyers", "url": "https://safelegalai.com/regulation/australia/qld-courts-generative-ai-guidelines-non-lawyers", "title": "The Use of Generative Artificial Intelligence: Guidelines for Responsible Use by Non-Lawyers", "body": "Queensland Courts", "bodyShort": "Queensland Courts", "jurisdiction": "australia", "region": "Queensland", "type": "guidance", "status": "in-force", "appliesTo": [ "litigants" ], "requirements": [ "verification", "confidentiality" ], "effectiveDate": "2024-05-01T00:00:00.000Z", "versions": [ { "version": "May 2024", "date": "2024-05-01T00:00:00.000Z", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/867502/artificial-intelligence-guidelines-for-non-lawyers.pdf" } ], "summary": "Guidelines published by Queensland Courts in May 2024 for self-represented litigants using generative AI. They caution that AI chatbots can produce inaccurate or fabricated legal content, that users must verify everything against official sources before relying on it in court, and that private or confidential information should not be entered into public tools.", "keyProvisions": [ "Generative AI output may contain fabricated cases and must be verified before use in court.", "Self-represented litigants remain responsible for the accuracy of what they file.", "Confidential or private information should not be entered into public AI chatbots.", "Companion guidelines address judicial officers' use of AI." ], "relatedIncidents": [], "sources": [ { "label": "Queensland Courts — AI guidelines for non-lawyers (PDF)", "url": "https://www.courts.qld.gov.au/__data/assets/pdf_file/0010/867502/artificial-intelligence-guidelines-for-non-lawyers.pdf" }, { "label": "Queensland Courts — using generative AI page", "url": "https://www.courts.qld.gov.au/going-to-court/prepare-for-court/representing-yourself/using-generative-ai" } ], "country": "AU", "categories": [ "litigant-in-person", "verification-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "vic-supreme-court-litigant-ai-guidelines-2024", "url": "https://safelegalai.com/regulation/australia/vic-supreme-court-litigant-ai-guidelines-2024", "title": "Guidelines for Litigants: Responsible Use of Artificial Intelligence in Litigation", "body": "Supreme Court of Victoria", "bodyShort": "Supreme Court of Victoria", "jurisdiction": "australia", "region": "Victoria", "type": "guidance", "status": "superseded", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "disclosure" ], "effectiveDate": "2024-05-01T00:00:00.000Z", "versions": [ { "version": "May 2024", "date": "2024-05-01T00:00:00.000Z", "url": "https://www.supremecourt.vic.gov.au/forms-fees-and-services/forms-templates-and-guidelines/guideline-responsible-use-of-ai-in-litigation" } ], "summary": "Guidelines issued by the Supreme Court of Victoria in May 2024 for litigants and practitioners using AI in litigation. They asked parties to understand the tools they used, verify output, and disclose AI use where it was material to a document before the court. They were replaced on 14 May 2026 by Practice Note SC GEN 25.", "keyProvisions": [ "Parties should understand how any AI tool works before relying on it.", "AI-assisted content must be checked for accuracy, including citations.", "Where AI has been used in preparing material, that use should be disclosed where relevant.", "Superseded by Practice Note SC GEN 25 on 14 May 2026." ], "relatedIncidents": [], "sources": [ { "label": "Supreme Court of Victoria — guideline on responsible use of AI in litigation", "url": "https://www.supremecourt.vic.gov.au/forms-fees-and-services/forms-templates-and-guidelines/guideline-responsible-use-of-ai-in-litigation" } ], "country": "AU", "categories": [ "litigant-in-person", "verification-duty", "disclosure-filings" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "int-svamc-ai-arbitration-guidelines", "url": "https://safelegalai.com/regulation/international/int-svamc-ai-arbitration-guidelines", "title": "SVAMC Guidelines on the Use of Artificial Intelligence in Arbitration (First Edition)", "body": "Silicon Valley Arbitration and Mediation Center", "bodyShort": "SVAMC", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "parties", "lawyers", "judges" ], "requirements": [ "disclosure", "confidentiality", "verification", "competence", "permissive" ], "effectiveDate": "2024-04-30T00:00:00.000Z", "versions": [ { "version": "First edition", "date": "2024-04-30T00:00:00.000Z", "url": "https://svamc.org/wp-content/uploads/SVAMC-AI-Guidelines-First-Edition.pdf" } ], "summary": "Published 30 April 2024, the SVAMC Guidelines are seven best-practice guidelines for parties, counsel and arbitrators using AI in international arbitration, adoptable by agreement or procedural order. They cover understanding tools, confidentiality, case-by-case disclosure, counsel's duty of competence and to verify output, AI-generated evidence, and a ban on arbitrators delegating their mandate.", "keyProvisions": [ "Guideline 3: disclosure of AI use is not required as a general matter and is decided case by case.", "Guideline 4: party representatives must observe competence and diligence rules and review AI output for accuracy.", "Guideline 5: parties should identify AI-generated material and its limitations.", "Guideline 6: an arbitrator shall not delegate any part of their personal mandate, especially decision-making, to an AI tool.", "Model clause for incorporating the Guidelines into arbitration agreements." ], "relatedIncidents": [], "sources": [ { "label": "SVAMC — Guidelines (PDF)", "url": "https://svamc.org/wp-content/uploads/SVAMC-AI-Guidelines-First-Edition.pdf" }, { "label": "SVAMC — publication announcement", "url": "https://svamc.org/svamc-publishes-guidelines-on-the-use-of-artificial-intelligence-in-arbitration/" } ], "country": "INT", "categories": [ "arbitration-adr", "disclosure-filings", "verification-duty", "confidentiality-client-data", "evidence-admissibility", "ai-decision-prohibition", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-mo-informal-opinion-2024-11", "url": "https://safelegalai.com/regulation/us-state/us-mo-informal-opinion-2024-11", "title": "Missouri Informal Opinion 2024-11: lawyer's use of generative artificial intelligence", "body": "Missouri Office of Legal Ethics Counsel", "bodyShort": "Missouri OLEC", "jurisdiction": "us-state", "region": "Missouri", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "verification", "supervision" ], "versions": [ { "version": "Informal Opinion 2024-11", "date": "2024-04-25T00:00:00.000Z", "url": "https://mo-legal-ethics.org/informal-opinion/2024-11/", "changes": "Adopted by the Office of Legal Ethics Counsel" } ], "summary": "Missouri's ethics counsel answered a firm asking how to build a generative-AI policy. The opinion walks through competence and training, platform assessment and confidentiality safeguards, obedience to court AI rules, professional independence, verification of AI output as nonlawyer work, supervision, fee reasonableness and candour toward the tribunal.", "keyProvisions": [ "Lawyers must obtain education and training to determine which generative AI is appropriate for firm use (Rule 4-1.1 and Comment [6])", "Before use, assess the platform's terms and conditions, the security of information input, how that information is used, and what data sources produce responses (Rule 4-1.6)", "Make reasonable efforts to prevent inadvertent or unauthorised disclosure or access, including by supervised persons (Rule 4-1.6(c), Comments [15] and [16])", "Rule 4-3.4(c) prohibits knowingly disobeying an obligation under a tribunal's rules, which reaches court AI orders", "Lawyers must maintain professional independence and may not rely solely on AI-generated content (Rule 4-5.4)", "AI-assisted work must be verified for accuracy and content as with any nonlawyer assistance (Rule 4-5.3)", "Managers and supervisors must ensure conformity and provide appropriate training (Rule 4-5.1, 4-5.3)", "Consider how generative AI affects the reasonableness of fees (Rule 4-1.5(a))", "Generative AI is not always accurate, so competence and supervision are needed to avoid false statements of material fact or law to a tribunal (Rule 4-3.3)" ], "relatedIncidents": [], "sources": [ { "label": "Missouri Informal Opinion 2024-11", "url": "https://mo-legal-ethics.org/informal-opinion/2024-11/" }, { "label": "Missouri informal opinions index", "url": "https://mo-legal-ethics.org/informal-opinions-search/by-year/" } ], "country": "US", "categories": [ "bar-guidance", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "uspto-ai-tools-practice-guidance", "url": "https://safelegalai.com/regulation/us-federal/uspto-ai-tools-practice-guidance", "title": "Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the USPTO", "body": "United States Patent and Trademark Office", "bodyShort": "USPTO", "jurisdiction": "us-federal", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "parties" ], "requirements": [ "verification", "confidentiality", "competence" ], "effectiveDate": "2024-04-11T00:00:00.000Z", "versions": [ { "version": "Federal Register notice (89 FR 25609)", "date": "2024-04-11T00:00:00.000Z", "url": "https://www.federalregister.gov/documents/2024/04/11/2024-07629/guidance-on-use-of-artificial-intelligence-based-tools-in-practice-before-the-united-states-patent" } ], "summary": "Guidance published by the United States Patent and Trademark Office on 11 April 2024 reminding practitioners that existing signature, candor and confidentiality duties apply when AI tools are used to prepare filings. Practitioners must review AI-assisted documents before submission, must not disclose client information to tools without safeguards, and remain responsible for accuracy.", "keyProvisions": [ "Signature requirements under 37 CFR 11.18 apply to AI-drafted submissions; the signer certifies reasonable inquiry.", "Duty of candor and good faith extends to AI-assisted prior-art searches and responses.", "Confidential client and export-controlled information must not be entered into tools that may disclose it.", "No duty to disclose AI use unless specifically required or material to patentability." ], "relatedIncidents": [], "sources": [ { "label": "Federal Register — USPTO AI tools guidance", "url": "https://www.federalregister.gov/documents/2024/04/11/2024-07629/guidance-on-use-of-artificial-intelligence-based-tools-in-practice-before-the-united-states-patent" } ], "country": "US", "categories": [ "verification-duty", "confidentiality-client-data", "competence-duty", "bar-guidance" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "ca-mb-law-society-generative-ai-guideline", "url": "https://safelegalai.com/regulation/canada/ca-mb-law-society-generative-ai-guideline", "title": "Generative Artificial Intelligence: Guideline for Use in the Practice of Law", "body": "Law Society of Manitoba", "bodyShort": "Law Society of Manitoba", "jurisdiction": "canada", "region": "Manitoba", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence", "supervision" ], "versions": [ { "version": "April 2024", "date": "2024-04-01T00:00:00.000Z", "url": "https://educationcentre.lawsociety.mb.ca/wp-content/uploads/sites/2/2024/04/Generative-Artificial-Intelligence-Guidelines-for-Use-in-the-Practice-of-Law.pdf", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the law society's published guideline listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A guideline published by the Law Society of Manitoba in April 2024 applying the Code of Professional Conduct to generative AI use, covering competence, confidentiality, verification of output and supervision, and complementing the Court of King's Bench disclosure practice direction.", "keyProvisions": [ "Lawyers remain responsible for AI-assisted work product.", "Confidential client information requires safeguards before entry into AI tools.", "Court practice directions requiring disclosure of AI use must be complied with." ], "relatedIncidents": [], "sources": [ { "label": "Law Society of Manitoba — generative AI guideline (PDF)", "url": "https://educationcentre.lawsociety.mb.ca/wp-content/uploads/sites/2/2024/04/Generative-Artificial-Intelligence-Guidelines-for-Use-in-the-Practice-of-Law.pdf" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "lso-licensee-generative-ai-white-paper", "url": "https://safelegalai.com/regulation/canada/lso-licensee-generative-ai-white-paper", "title": "White Paper: Licensee Use of Generative Artificial Intelligence", "body": "Law Society of Ontario", "bodyShort": "LSO", "jurisdiction": "canada", "region": "Ontario", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence", "supervision" ], "effectiveDate": "2024-04-01T00:00:00.000Z", "versions": [ { "version": "April 2024", "date": "2024-04-01T00:00:00.000Z", "url": "https://lawsocietyontario-dwd0dscmayfwh7bj.a01.azurefd.net/media/lso/media/lawyers/practice-supports-resources/white-paper-on-licensee-use-of-generative-artificial-intelligence-en.pdf", "archiveUrl": "https://web.archive.org/web/20260830190840/https://lawsocietyontario-dwd0dscmayfwh7bj.a01.azurefd.net/media/lso/media/lawyers/practice-supports-resources/white-paper-on-licensee-use-of-generative-artificial-intelligence-en.pdf" } ], "summary": "The Law Society of Ontario's April 2024 white paper on licensee use of generative AI applies the Rules of Professional Conduct and Paralegal Rules to lawyers and paralegals in Ontario. It is permissive with cautions: licensees must verify output, protect confidentiality and privilege, remain competent, and need not disclose AI use to clients solely because AI was used.", "keyProvisions": [ "Licensees are responsible for all work product regardless of AI assistance.", "AI output, including case citations, must be verified before reliance.", "No obligation to disclose AI use to a client solely because AI was used, though disclosure may be appropriate.", "Confidential and privileged information requires safeguards before entry into AI tools." ], "relatedIncidents": [ "ko-v-li", "r-v-chand" ], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Law Society of Ontario — white paper (PDF)", "url": "https://lawsocietyontario-dwd0dscmayfwh7bj.a01.azurefd.net/media/lso/media/lawyers/practice-supports-resources/white-paper-on-licensee-use-of-generative-artificial-intelligence-en.pdf" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-dc-bar-ethics-opinion-388", "url": "https://safelegalai.com/regulation/us-state/us-dc-bar-ethics-opinion-388", "title": "D.C. Bar Legal Ethics Opinion 388: Attorneys' Use of Generative Artificial Intelligence in Client Matters", "body": "D.C. Bar Legal Ethics Committee", "bodyShort": "D.C. Bar", "jurisdiction": "us-state", "region": "District of Columbia", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "verification", "confidentiality", "consent", "supervision" ], "versions": [ { "version": "Opinion 388", "date": "2024-04-01T00:00:00.000Z", "url": "https://www.dcbar.org/for-lawyers/legal-ethics/ethics-opinions-210-present/ethics-opinion-388", "changes": "Published April 2024; the source gives a month only (published April 2024), so the day is recorded as the 1st" } ], "summary": "The D.C. Bar's generative-AI opinion. Lawyers must understand that GAI generates rather than retrieves, must not rely blindly on output, must check whether a product retains inputs or feeds them to other users before entering client confidences, may bill only time actually spent, and must supervise AI use by others under Rules 5.1 and 5.3.", "keyProvisions": [ "Competence includes understanding enough about the technology to be reasonably confident it will advance the client's interests (Rule 1.1)", "Generative AI is not a search engine over a database; lawyers who rely blindly on GAI output do so at considerable peril", "Before entering client confidences, determine whether the product saves the information and whether the interaction will affect answers given to future users outside the firm (Rule 1.6)", "Where a fee is based solely on time spent, a lawyer may bill only for time actually spent even if GAI reduces it (Rule 1.5)", "Absent prior agreement, a lawyer may not charge separately for the perceived value of GAI work but may pass through out-of-pocket GAI expenses the client agreed to pay", "Lawyers must be attentive to candour and fairness duties, especially before tribunals that have adopted AI rules or orders (Rules 3.3, 3.4)", "Supervisory duties under Rules 5.1 and 5.3 extend to supervised lawyers' and nonlawyers' use of GAI" ], "relatedIncidents": [], "sources": [ { "label": "Ethics Opinion 388", "url": "https://www.dcbar.org/for-lawyers/legal-ethics/ethics-opinions-210-present/ethics-opinion-388" } ], "country": "US", "categories": [ "bar-guidance", "fees-billing", "confidentiality-client-data", "verification-duty", "competence-duty", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ky-bar-ethics-opinion-e-457", "url": "https://safelegalai.com/regulation/us-state/ky-bar-ethics-opinion-e-457", "title": "KBA E-457: The Ethical Use of Artificial Intelligence in the Practice of Law", "body": "Kentucky Bar Association Ethics Committee", "bodyShort": "Kentucky Bar Association", "jurisdiction": "us-state", "region": "Kentucky", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "competence", "confidentiality", "verification", "supervision" ], "effectiveDate": "2024-03-15T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-03-15T00:00:00.000Z", "url": "https://kybar.org/Portals/0/Admin/Ethics%20Opinions/KBA%20E-457.pdf" } ], "summary": "Kentucky Bar Association Ethics Opinion E-457, issued 15 March 2024, addresses generative AI under the Kentucky Rules of Professional Conduct. Lawyers may use AI but must be competent in the technology, verify its output, protect client confidences, and supervise staff. There is no mandatory duty to disclose AI use to clients unless the client asks or the use is material.", "keyProvisions": [ "Lawyers must understand the AI tools they use and their risks.", "No mandatory disclosure of AI use to clients absent a client request or material impact on the representation.", "AI output must be verified before use in client work or court filings.", "Confidential information requires safeguards before entry into third-party tools.", "Supervisors are responsible for AI use by subordinate lawyers and nonlawyers." ], "relatedIncidents": [], "sources": [ { "label": "Kentucky Bar Association — KBA E-457 (PDF)", "url": "https://kybar.org/Portals/0/Admin/Ethics%20Opinions/KBA%20E-457.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "sg-pdpc-advisory-guidelines-ai-recommendation-decision-systems", "url": "https://safelegalai.com/regulation/singapore/sg-pdpc-advisory-guidelines-ai-recommendation-decision-systems", "title": "Advisory Guidelines on Use of Personal Data in AI Recommendation and Decision Systems", "body": "Personal Data Protection Commission, Singapore", "bodyShort": "PDPC", "jurisdiction": "singapore", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "consent", "confidentiality", "record-keeping" ], "effectiveDate": "2024-03-01T00:00:00.000Z", "versions": [ { "version": "Issued", "date": "2024-03-01T00:00:00.000Z", "url": "https://www.pdpc.gov.sg/-/media/files/pdpc/pdf-files/advisory-guidelines/advisory-guidelines-on-the-use-of-personal-data-in-ai-recommendation-and-decision-systems.pdf" } ], "summary": "PDPC advisory guidelines issued 1 March 2024 explaining how the Personal Data Protection Act applies to the collection and use of personal data in developing, testing, monitoring and deploying AI recommendation and decision systems, including the business improvement and research exceptions and the obligations of organisations that procure such systems.", "keyProvisions": [ "The PDPA applies to personal data used to design, train and deploy AI systems.", "Sets out when the business improvement exception and the research exception permit use of personal data in AI development.", "Addresses notification and consent for collection and use of personal data at deployment.", "Addresses obligations of organisations that procure AI systems processing personal data." ], "relatedIncidents": [], "sources": [ { "label": "PDPC — Advisory Guidelines (official PDF)", "url": "https://www.pdpc.gov.sg/-/media/files/pdpc/pdf-files/advisory-guidelines/advisory-guidelines-on-the-use-of-personal-data-in-ai-recommendation-and-decision-systems.pdf" } ], "country": "SG", "categories": [ "data-protection", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-sk-law-society-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/canada/ca-sk-law-society-generative-ai-guidelines", "title": "Guidelines for the Use of Generative Artificial Intelligence in the Practice of Law", "body": "Law Society of Saskatchewan", "bodyShort": "Law Society of Saskatchewan", "jurisdiction": "canada", "region": "Saskatchewan", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence", "supervision" ], "versions": [ { "version": "February 2024", "date": "2024-02-01T00:00:00.000Z", "url": "https://www.lawsociety.sk.ca/wp-content/uploads/Law-Society-of-Saskatchewan-Generative-Artificial-Intelligence-Guidelines.pdf", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the law society's published guidelines listing; the PDF was not read in full, so this record is unverified." } ], "summary": "Guidelines published by the Law Society of Saskatchewan in February 2024 applying the Code of Professional Conduct to generative artificial intelligence: competence in the tool being used, confidentiality of client information, verification of output before reliance, and supervision of AI-assisted work carried out by staff and students.", "keyProvisions": [ "Lawyers must understand the tools they use and remain competent in their limits.", "Client confidential information requires safeguards before entry into AI tools.", "AI output, including citations, must be verified before reliance." ], "relatedIncidents": [], "sources": [ { "label": "Law Society of Saskatchewan — generative AI guidelines (PDF)", "url": "https://www.lawsociety.sk.ca/wp-content/uploads/Law-Society-of-Saskatchewan-Generative-Artificial-Intelligence-Guidelines.pdf" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty", "supervision" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "cepej-information-note-generative-ai", "url": "https://safelegalai.com/regulation/international/cepej-information-note-generative-ai", "title": "Information note on the use of generative AI by judicial professionals", "body": "European Commission for the Efficiency of Justice, Council of Europe", "bodyShort": "CEPEJ", "jurisdiction": "international", "type": "guidance", "status": "superseded", "appliesTo": [ "judges", "court-staff", "lawyers" ], "requirements": [ "verification", "confidentiality" ], "effectiveDate": "2024-02-01T00:00:00.000Z", "versions": [ { "version": "February 2024", "date": "2024-02-01T00:00:00.000Z", "url": "https://www.coe.int/en/web/cepej/newpage-cyber" } ], "summary": "An information note issued by the Council of Europe's CEPEJ in February 2024 cautioning judges, court staff and lawyers about generative AI. It highlighted hallucinated output, confidentiality of case data and the need for human oversight. The note was superseded on 19 December 2025 by CEPEJ's formal guidelines on generative AI in courts.", "keyProvisions": [ "Warned that generative AI can produce plausible but false legal content.", "Advised against entering confidential case information into public tools.", "Recommended that any judicial use remain under human control and be verified." ], "relatedIncidents": [], "sources": [ { "label": "Council of Europe — CEPEJ cyberjustice page", "url": "https://www.coe.int/en/web/cepej/newpage-cyber" } ], "country": "INT", "categories": [ "judicial-use", "verification-duty", "confidentiality-client-data", "court-own-use", "bar-guidance" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ct-judicial-branch-responsible-ai-policy-framework", "url": "https://safelegalai.com/regulation/us-state/ct-judicial-branch-responsible-ai-policy-framework", "title": "Responsible AI Policy Framework v2.1", "body": "Connecticut Judicial Branch", "bodyShort": "Connecticut Judicial Branch", "jurisdiction": "us-state", "region": "Connecticut", "type": "policy", "status": "in-force", "appliesTo": [ "court-staff", "judges" ], "requirements": [ "risk-classification", "record-keeping", "competence" ], "effectiveDate": "2024-02-01T00:00:00.000Z", "versions": [ { "version": "v2.1", "date": "2024-02-01T00:00:00.000Z", "url": "https://www.jud.ct.gov/faq/CTJBResponsibleAIPolicyFramework2.1.24.pdf" } ], "summary": "An internal governance framework adopted by the Connecticut Judicial Branch on 1 February 2024 for how the branch itself procures, deploys and monitors AI systems. It sets principles of accountability, transparency, fairness and security for court technology, requires risk assessment before deployment, and keeps human decision-makers responsible. It does not regulate filings by lawyers or litigants.", "keyProvisions": [ "AI systems used by the branch must be inventoried and risk-assessed before deployment.", "Human oversight is required for any AI that affects court operations or the public.", "Procurement must address data security, bias testing and vendor accountability." ], "relatedIncidents": [], "sources": [ { "label": "Connecticut Judicial Branch — Responsible AI Policy Framework v2.1 (PDF)", "url": "https://www.jud.ct.gov/faq/CTJBResponsibleAIPolicyFramework2.1.24.pdf" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "nj-supreme-court-preliminary-ai-guidelines", "url": "https://safelegalai.com/regulation/us-state/nj-supreme-court-preliminary-ai-guidelines", "title": "Preliminary Guidelines on the Use of Artificial Intelligence by New Jersey Lawyers", "body": "Supreme Court of New Jersey (Notice to the Bar)", "bodyShort": "New Jersey Supreme Court", "jurisdiction": "us-state", "region": "New Jersey", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "effectiveDate": "2024-01-24T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-01-24T00:00:00.000Z", "url": "https://www.njcourts.gov/notices" } ], "summary": "Preliminary guidelines issued by the Supreme Court of New Jersey on 24 January 2024 by Notice to the Bar. Lawyers in New Jersey may use AI, and there is no blanket duty to disclose that use, but the existing Rules of Professional Conduct on competence, candor, confidentiality, supervision and communication apply in full and require verification of AI output.", "keyProvisions": [ "No general obligation to disclose AI use to clients or courts absent a specific rule.", "Lawyers must verify the accuracy of AI-generated research and drafting.", "Confidential client information must be protected when using AI tools.", "Supervisory lawyers must ensure subordinate use of AI complies with the RPCs." ], "relatedIncidents": [], "sources": [ { "label": "New Jersey Courts — notices to the bar", "url": "https://www.njcourts.gov/notices" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "us-la-supreme-court-letter-ai", "url": "https://safelegalai.com/regulation/us-state/us-la-supreme-court-letter-ai", "title": "Louisiana Supreme Court letter to the Louisiana State Bar Association on the emergence of artificial intelligence", "body": "Supreme Court of Louisiana, Office of the General Counsel", "bodyShort": "Louisiana Supreme Court", "jurisdiction": "us-state", "region": "Louisiana", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "judges" ], "requirements": [ "competence", "verification", "confidentiality" ], "versions": [ { "version": "Original", "date": "2024-01-22T00:00:00.000Z", "url": "https://www.lsba.org/documents/News/LSBANews/LASCLetterAI.pdf", "changes": "Letter from the Court's General Counsel on behalf of the Justices" } ], "summary": "The Louisiana Supreme Court chose not to make an AI rule. Its January 2024 letter to the State Bar states existing ethical and professional rules are broad enough for AI, lists the conduct rules and Code of Civil Procedure articles that apply, confirms judges' parallel responsibilities, and notes the Court's Technology Commission and Rules Committee will monitor developments.", "keyProvisions": [ "The ethical and professional rules governing the bench and the Bar are robust and broad enough to cover the landscape of issues presented by AI in its current forms", "Attorneys remain ultimately responsible for their work product and pleadings regardless of AI use, and must maintain technology competence, protect confidential client information and avoid misrepresenting fact or law", "Cites Rules of Professional Conduct 1.1, 1.3, 1.4, 1.6, 1.15, 3.1, 3.3 and 5.3 and Code of Civil Procedure articles 371 and 863", "Judges remain ultimately responsible for their opinions and decisions and must maintain competence in technology (Canons 1, 2A, 2B, 3A(1), 3A(4), 3A(7), 3B(1))", "The Court's Technology Commission and Rules Committee will monitor whether future rule changes are needed" ], "relatedIncidents": [], "sources": [ { "label": "Louisiana Supreme Court letter (PDF)", "url": "https://www.lsba.org/documents/News/LSBANews/LASCLetterAI.pdf" } ], "country": "US", "categories": [ "bar-guidance", "judicial-use", "disclosure-filings", "confidentiality-client-data", "verification-duty", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "fl-bar-ethics-opinion-24-1", "url": "https://safelegalai.com/regulation/us-state/fl-bar-ethics-opinion-24-1", "title": "Florida Bar Ethics Opinion 24-1: Lawyers' use of generative artificial intelligence", "body": "The Florida Bar", "bodyShort": "Florida Bar", "jurisdiction": "us-state", "region": "Florida", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "consent", "confidentiality", "verification", "supervision", "competence" ], "effectiveDate": "2024-01-19T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2024-01-19T00:00:00.000Z", "url": "https://www.floridabar.org/etopinions/opinion-24-1/" } ], "summary": "Ethics Opinion 24-1, issued by the Florida Bar on 19 January 2024, permits Florida lawyers to use generative AI under the Rules Regulating the Florida Bar. Lawyers must obtain informed client consent before disclosing confidential information to a third-party AI tool, must verify output, must supervise the tool like nonlawyer staff, and may not charge for AI time savings.", "keyProvisions": [ "Informed client consent is required before confidential information is disclosed to a third-party generative AI.", "Lawyers must review all AI-generated work product for accuracy and completeness.", "AI tools must be supervised in the same way as nonlawyer assistants.", "Fees must be reasonable; lawyers may not bill clients for time saved by AI.", "AI use in advertising and intake chatbots must not be misleading." ], "relatedIncidents": [ "gleason-v-marcus" ], "sources": [ { "label": "The Florida Bar — Ethics Opinion 24-1", "url": "https://www.floridabar.org/etopinions/opinion-24-1/" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "sa-sdaia-generative-ai-guidelines-government", "url": "https://safelegalai.com/regulation/gulf/sa-sdaia-generative-ai-guidelines-government", "title": "Generative Artificial Intelligence Guidelines for Government", "body": "Saudi Data and Artificial Intelligence Authority", "bodyShort": "SDAIA", "jurisdiction": "gulf", "type": "guidance", "status": "in-force", "appliesTo": [ "court-staff", "all" ], "requirements": [ "verification", "supervision", "confidentiality", "risk-classification" ], "versions": [ { "version": "First issue", "date": "2024-01-11T00:00:00.000Z", "url": "https://sdaia.gov.sa/en/SDAIA/about/Files/GenAIGuidelinesForGovernmentENCompressed.pdf", "changes": "Issued January 2024; a 2025 edition is published at the same URL. The PDF was not read in full, so this record is unverified." } ], "summary": "Non-binding SDAIA guidance for Saudi government entities and employees on adopting and overseeing generative AI. It sets principles, risk classifications, data-handling rules and role definitions with a compliance checklist aligned to the Personal Data Protection Law, cybersecurity rules and the AI Ethics Principles. It reaches the Ministry of Justice and the courts as government bodies.", "keyProvisions": [ "Requires transparency, accountability, privacy protection, human oversight and risk management in government generative-AI use.", "Sets risk classifications and role definitions for government users of generative AI.", "Includes a compliance checklist aligned with the PDPL, national cybersecurity controls and the AI Ethics Principles.", "Non-binding but treated as the authoritative baseline for Saudi public-sector AI adoption." ], "relatedIncidents": [], "sources": [ { "label": "SDAIA — Generative AI Guidelines for Government (official PDF)", "url": "https://sdaia.gov.sa/en/SDAIA/about/Files/GenAIGuidelinesForGovernmentENCompressed.pdf" } ], "country": "SA", "categories": [ "court-own-use", "verification-duty", "confidentiality-client-data", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "sa-sdaia-generative-ai-guidelines-public", "url": "https://safelegalai.com/regulation/gulf/sa-sdaia-generative-ai-guidelines-public", "title": "Generative Artificial Intelligence Guidelines for the Public", "body": "Saudi Data and Artificial Intelligence Authority", "bodyShort": "SDAIA", "jurisdiction": "gulf", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "litigants", "providers" ], "requirements": [ "verification", "disclosure", "supervision" ], "versions": [ { "version": "First issue", "date": "2024-01-11T00:00:00.000Z", "url": "https://sdaia.gov.sa/en/SDAIA/about/Files/GenerativeAIPublicEN.pdf", "changes": "Issued January 2024; a 2025 edition is published at the same URL. The PDF was not read in full, so this record is unverified." } ], "summary": "Non-binding SDAIA guidance for individuals, organisations and developers using generative AI in Saudi Arabia. It identifies key harms — misinformation, deepfakes, privacy intrusion, biased output and cybersecurity risk — and recommends disclosure of synthetic content, provenance metadata, human-in-the-loop controls and data governance. It is not addressed to legal practice.", "keyProvisions": [ "Recommends disclosure of synthetic content and use of provenance metadata.", "Recommends human-in-the-loop controls and verification of generative output.", "Identifies misinformation, deepfakes, privacy intrusion, bias and cybersecurity as the principal risks." ], "relatedIncidents": [], "sources": [ { "label": "SDAIA — Generative AI Guidelines for the Public (official PDF)", "url": "https://sdaia.gov.sa/en/SDAIA/about/Files/GenerativeAIPublicEN.pdf" } ], "country": "SA", "categories": [ "verification-duty", "ai-act-horizontal", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "fi-bar-tietoturvaohje-b051", "url": "https://safelegalai.com/regulation/other/fi-bar-tietoturvaohje-b051", "title": "B 05.1 Data Security Guideline [Tietoturvaohje]", "body": "Finnish Bar Association [Suomen Asianajajaliitto / Suomen Asianajajat]", "bodyShort": "Finnish Bar Association", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "competence", "consent", "supervision", "record-keeping" ], "versions": [ { "version": "9.6.2023 amendment", "date": "2024-01-01T00:00:00.000Z", "url": "https://asianajajat.fi/wp-content/uploads/2025/01/B-05.1-Tietoturvaohje-24.1.2019-muut.-16.1.2020-ja-9.6.2023-voimassa-1.1.2024-alkaen.pdf", "changes": "Delegation amendments of 9 June 2023 in force 1 Jan 2024: added competence and documented-training duty, security policy and external audit for firms of 10+ staff, business-grade software requirement, device management and access control, encryption, backup and supplier-contract terms." }, { "version": "16.1.2020 amendment", "date": "2020-02-01T00:00:00.000Z", "url": "https://asianajajat.fi/wp-content/uploads/2025/01/B-05.1-Tietoturvaohje-24.1.2019-muut.-16.1.2020-ja-9.6.2023-voimassa-1.1.2024-alkaen.pdf", "changes": "Added point 4 barring audits or information requests that would collect or disclose data on client relationships or engagements to outside parties." }, { "version": "Original", "date": "2019-06-01T00:00:00.000Z", "url": "https://asianajajat.fi/wp-content/uploads/2025/01/B-05.1-Tietoturvaohje-24.1.2019-muut.-16.1.2020-ja-9.6.2023-voimassa-1.1.2024-alkaen.pdf", "changes": "Guideline adopted by the Bar's delegation on 24 January 2019, in force 1 June 2019." } ], "summary": "The Finnish Bar's binding information-security guideline, adopted by its delegation and in force in its current form since 1 January 2024. Eighteen numbered duties on advocates covering competence and documented training, security policy, external audits, business-grade software, device and access management, encryption, backups, supplier contracts and secure disposal.", "keyProvisions": [ "Point 6: software and services used in legal practice must be intended for business use; other services may be used only with the client's consent — the hook the Bar applies to AI applications.", "Point 1: the advocate must ensure their own and staff information-security competence is sufficiently high, assured by planned training, with evidence produced on request as for continuing education (B 9).", "Point 14: contracts with all external service providers must meet security requirements, in particular confidentiality terms, scoped access rights with periodic review, and agreed data destruction and portability.", "Points 2 and 3: firms of 10 or more employees must have a board-approved security policy and commission periodic external security audits, logged, and after significant system changes.", "Point 8: devices, tables, databases and cloud services holding client data must be encrypted; devices no longer receiving updates must be retired." ], "relatedIncidents": [], "sources": [ { "label": "Suomen Asianajajat — B 05.1 Tietoturvaohje (PDF)", "url": "https://asianajajat.fi/wp-content/uploads/2025/01/B-05.1-Tietoturvaohje-24.1.2019-muut.-16.1.2020-ja-9.6.2023-voimassa-1.1.2024-alkaen.pdf" }, { "label": "Suomen Asianajajat — regulations and guidelines index", "url": "https://asianajajat.fi/asianajajalle/saantely/saadoksia-ja-ohjeita/" } ], "country": "FI", "categories": [ "confidentiality-client-data", "competence-duty", "client-consent-transparency", "procurement-vendor", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ae-difc-courts-practical-guidance-note-2-2023-ai", "url": "https://safelegalai.com/regulation/gulf/ae-difc-courts-practical-guidance-note-2-2023-ai", "title": "Practical Guidance Note No. 2 of 2023: large language models and generative AI in DIFC Courts proceedings", "body": "Dubai International Financial Centre Courts", "bodyShort": "DIFC Courts", "jurisdiction": "gulf", "region": "Dubai International Financial Centre", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "verification", "confidentiality" ], "versions": [ { "version": "Original", "date": "2023-12-21T00:00:00.000Z", "url": "https://www.difccourts.ae/rules-decisions/practice-directions/practical-guidance-note-no-2-2023-guidelines-use-large-language-models-and-generative-ai-proceedings-difc-courts", "changes": "First issue." } ], "summary": "Guidance issued by the DIFC Courts on 21 December 2023 for parties and practitioners using large language models and generative AI in proceedings. It sets out four principles — transparency, accuracy and reliability, professional and legal obligations, and avoiding over-reliance — and requires parties to declare AI-generated content at the earliest possible opportunity.", "keyProvisions": [ "All parties and the Courts should be made aware of any AI-generated content used during the process.", "Parties must declare at the earliest opportunity that they have used or intend to use AI-generated content; late disclosure risks adjournment and lost trial dates.", "AI-generated material must be verified for accuracy and reliability against independent sources such as case law and statutes before submission.", "Practitioners must comply with the DIFC Courts Mandatory Code of Conduct and applicable data-protection and intellectual-property law.", "Technology must assist submissions, not replace the integral human decision-making required.", "Witness statements must remain in the intended witness's own words under RDC 29.24." ], "relatedIncidents": [], "sources": [ { "label": "DIFC Courts — Practical Guidance Note No. 2 of 2023 (official)", "url": "https://www.difccourts.ae/rules-decisions/practice-directions/practical-guidance-note-no-2-2023-guidelines-use-large-language-models-and-generative-ai-proceedings-difc-courts" } ], "country": "AE", "categories": [ "disclosure-filings", "verification-duty", "evidence-admissibility", "confidentiality-client-data", "bar-guidance" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "es-rdl-6-2023-actuaciones-automatizadas-asistidas", "url": "https://safelegalai.com/regulation/other/es-rdl-6-2023-actuaciones-automatizadas-asistidas", "title": "Royal Decree-Law 6/2023: automated, proactive and assisted actions in the Administration of Justice", "body": "Government of Spain (Gobierno de Espana)", "bodyShort": "Gobierno de Espana", "jurisdiction": "other", "type": "statute", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "permissive" ], "effectiveDate": "2023-12-20T00:00:00.000Z", "versions": [ { "version": "BOE-A-2023-25758, 20 December 2023 (consolidated)", "date": "2023-12-20T00:00:00.000Z", "url": "https://www.boe.es/buscar/act.php?id=BOE-A-2023-25758" } ], "summary": "Digital-efficiency measures for the Spanish justice system. It distinguishes automated actions of simple processing, proactive actions such as alerts, and assisted actions in which systems including AI generate full or partial drafts to support judges, prosecutors and court lawyers; a draft cannot become a resolution without the operator's intervention and validation.", "keyProvisions": [ "Arts 56-57: definitions of automated, proactive and assisted actions in the administration of justice.", "Assisted drafts serve only as support and require human validation before becoming resolutions.", "Art. 35: data-oriented justice enables quality draft generation." ], "relatedIncidents": [], "sources": [ { "label": "BOE consolidated text", "url": "https://www.boe.es/buscar/act.php?id=BOE-A-2023-25758" } ], "country": "ES", "categories": [ "court-own-use", "ai-decision-prohibition" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "id-komdigi-surat-edaran-9-2023-etika-ai", "url": "https://safelegalai.com/regulation/other/id-komdigi-surat-edaran-9-2023-etika-ai", "title": "Circular Letter No. 9 of 2023 on Artificial Intelligence Ethics [SE Menkominfo No. 9/2023 Etika Kecerdasan Artifisial]", "body": "Ministry of Communication and Digital [Kementerian Komunikasi dan Digital, Komdigi]", "bodyShort": "Komdigi", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "providers", "all" ], "requirements": [ "permissive", "risk-classification", "confidentiality" ], "effectiveDate": "2023-12-19T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-12-19T00:00:00.000Z", "url": "https://jdih.komdigi.go.id/produk_hukum/view/id/883/t/surat-edaran-menteri-komunikasi-dan-informatika-nomor-9-tahun-2023" } ], "summary": "Indonesia's only in-force AI instrument as at September 2026: a ministerial circular of 19 December 2023 setting nine ethical values for artificial intelligence, addressed to AI businesses and to public and private electronic system operators. It makes no mention of courts, judges or legal services.", "keyProvisions": [ "Sets nine ethical values: inclusivity, humanity, security, accessibility, transparency, credibility and accountability, personal data protection, sustainability, and intellectual property.", "Addressed to AI businesses and public and private electronic system operators.", "Contains no provisions on courts, judges or legal services." ], "relatedIncidents": [], "sources": [ { "label": "JDIH Komdigi — SE Menteri No. 9 Tahun 2023", "url": "https://jdih.komdigi.go.id/produk_hukum/view/id/883/t/surat-edaran-menteri-komunikasi-dan-informatika-nomor-9-tahun-2023" } ], "country": "ID", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "nz-courts-generative-ai-guidelines", "url": "https://safelegalai.com/regulation/nz/nz-courts-generative-ai-guidelines", "title": "Guidelines for use of generative artificial intelligence in Courts and Tribunals", "body": "Courts of New Zealand (Chief Justice)", "bodyShort": "Courts of New Zealand", "jurisdiction": "nz", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "lawyers", "litigants", "court-staff" ], "requirements": [ "verification", "confidentiality", "permissive" ], "effectiveDate": "2023-12-07T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-12-07T00:00:00.000Z", "url": "https://www.courtsofnz.govt.nz/going-to-court/practice-directions/practice-guidelines/all-benches/guidelines-for-use-of-generative-artificial-intelligence-in-courts-and-tribunals" } ], "summary": "Three sets of guidelines issued by the Chief Justice of New Zealand on 7 December 2023 for judges, lawyers and non-lawyers using generative AI in courts and tribunals. Existing duties continue to apply, users must verify all AI output and protect confidential information, and there is no mandatory disclosure of AI use. The guidelines remain unchanged as of September 2026.", "keyProvisions": [ "Separate guidelines for judicial officers, lawyers and non-lawyers.", "AI output must be checked for accuracy; fabricated citations are the user's responsibility.", "Confidential, suppressed or private information must not be entered into public tools.", "No requirement to disclose AI use in court documents." ], "relatedIncidents": [ "jones-v-family-court-whangarei" ], "sources": [ { "label": "Courts of New Zealand — generative AI guidelines", "url": "https://www.courtsofnz.govt.nz/going-to-court/practice-directions/practice-guidelines/all-benches/guidelines-for-use-of-generative-artificial-intelligence-in-courts-and-tribunals" } ], "country": "NZ", "categories": [ "judicial-use", "verification-duty", "confidentiality-client-data", "litigant-in-person", "evidence-admissibility", "court-own-use", "supervision", "disclosure-filings" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "ca-nsbs-ai-practice-of-law-guide", "url": "https://safelegalai.com/regulation/canada/ca-nsbs-ai-practice-of-law-guide", "title": "Artificial Intelligence in the Practice of Law: What is AI and can or should I use it in my practice?", "body": "Nova Scotia Barristers' Society", "bodyShort": "NSBS", "jurisdiction": "canada", "region": "Nova Scotia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "versions": [ { "version": "December 2023", "date": "2023-12-01T00:00:00.000Z", "url": "https://nsbs.org/wp-content/uploads/2023/12/short-guide-on-A.I.pdf", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the society's published guide listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A short guide published by the Nova Scotia Barristers' Society in December 2023 explaining what generative AI is and whether lawyers can or should use it. It applies the Code of Professional Conduct to AI: competence, confidentiality, verification of output, and the lawyer's continuing responsibility for the work.", "keyProvisions": [ "Lawyers must understand what AI tools do before using them in practice.", "Client confidentiality must be protected when using AI tools.", "AI output must be verified; the lawyer remains responsible for the work." ], "relatedIncidents": [], "sources": [ { "label": "Nova Scotia Barristers' Society — short guide on AI (PDF)", "url": "https://nsbs.org/wp-content/uploads/2023/12/short-guide-on-A.I.pdf" }, { "label": "NSBS — Navigating the Frontier of Generative AI in Legal Practice", "url": "https://nsbs.org/society-news/legal-services-support/navigating-the-frontier-of-generative-ai-in-legal-practice/" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "int-ccje-opinion-26-assistive-technology", "url": "https://safelegalai.com/regulation/international/int-ccje-opinion-26-assistive-technology", "title": "CCJE Opinion No. 26 (2023): Moving forward - the use of assistive technology in the judiciary", "body": "Consultative Council of European Judges, Council of Europe", "bodyShort": "CCJE", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "competence", "permissive" ], "versions": [ { "version": "Final", "date": "2023-12-01T00:00:00.000Z", "url": "https://rm.coe.int/ccje-opinion-no-26-2023-final/1680adade7" } ], "summary": "Opinion of the Council of Europe's judges' advisory body, adopted at its 24th plenary in December 2023, on technology including AI in courts. It holds that technology must assist and not replace judicial decision-making, that judges should be involved in decisions to adopt tools, and that they should be kept up to date on technological innovation through training.", "keyProvisions": [ "Assistive technology, including AI, must support rather than replace the judge's decision-making.", "Judges should be consulted and, where necessary, concur in the adoption of new technology in courts.", "Judicial training must keep judges up to date with technological innovation." ], "relatedIncidents": [], "sources": [ { "label": "CCJE Opinion No. 26 (PDF)", "url": "https://rm.coe.int/ccje-opinion-no-26-2023-final/1680adade7" } ], "country": "INT", "categories": [ "judicial-use", "ai-decision-prohibition", "training-cle" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "my-bar-council-circular-342-2023-chatgpt", "url": "https://safelegalai.com/regulation/other/my-bar-council-circular-342-2023-chatgpt", "title": "Circular No. 342/2023: Risks and Precautions in Using Generative Artificial Intelligence in the Legal Profession", "body": "Bar Council Malaysia [Majlis Peguam]", "bodyShort": "Bar Council Malaysia", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "competence" ], "effectiveDate": "2023-11-24T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-11-24T00:00:00.000Z", "url": "https://www.malaysianbar.org.my/cms/upload_files/document/Circular%20No%20342-2023.pdf", "changes": "Supplemented by Circular No. 242/2025." } ], "summary": "The Bar Council's first AI circular, issued 24 November 2023 by its Technology, Cyber and Privacy Laws Committee, setting out the risks of ChatGPT in legal practice and the precautions practitioners should take. It remains in force alongside the 2025 circular that supplements it.", "keyProvisions": [ "Output must be independently verified against traditional legal databases.", "Sets out risks including fabricated citations, confidentiality exposure and over-reliance.", "Addressed specifically to the use of ChatGPT and comparable public chatbots by members of the profession." ], "relatedIncidents": [], "sources": [ { "label": "Bar Council Malaysia — Circular No. 342/2023 (official PDF)", "url": "https://www.malaysianbar.org.my/cms/upload_files/document/Circular%20No%20342-2023.pdf" } ], "country": "MY", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-co-people-v-crabill-discipline", "url": "https://safelegalai.com/regulation/us-state/us-co-people-v-crabill-discipline", "title": "People v. Zachariah C. Crabill, 23PDJ067 (Colo. O.P.D.J. 22 Nov 2023)", "body": "Colorado Office of the Presiding Disciplinary Judge", "bodyShort": "Colorado OPDJ", "jurisdiction": "us-state", "region": "Colorado", "type": "judgment-guidance", "status": "in-force", "appliesTo": [ "lawyers" ], "requirements": [ "verification", "competence" ], "versions": [ { "version": "Stipulation to discipline", "date": "2023-11-22T00:00:00.000Z", "url": "https://www.coloradolegalregulation.com/wp-content/uploads/PDJ/Decisions/Crabill,%20Stipulation%20to%20Discipline,%2023PDJ067,%2011-22-23.pdf", "changes": "Approved by the Presiding Disciplinary Judge" } ], "summary": "Colorado's Presiding Disciplinary Judge suspended Zachariah Crabill for one year and one day, with ninety days served, after he filed ChatGPT-generated citations without reading them, failed to correct the record and blamed a legal intern. Violations found of Colo. RPC 1.1, 1.3, 3.3(a)(1) and 8.4(c).", "keyProvisions": [ "Not reading or verifying AI-generated citations before filing violated Colo. RPC 1.1 and 1.3", "Failing to alert the court or withdraw the motion after discovering the citations were fictitious violated Colo. RPC 3.3(a)(1)", "Falsely attributing the errors to a legal intern violated Colo. RPC 8.4(c)", "Sanction: suspension of one year and one day, ninety days active, remainder stayed on two years' probation" ], "relatedIncidents": [], "sources": [ { "label": "Stipulation to discipline (PDF)", "url": "https://www.coloradolegalregulation.com/wp-content/uploads/PDJ/Decisions/Crabill,%20Stipulation%20to%20Discipline,%2023PDJ067,%2011-22-23.pdf" } ], "country": "US", "categories": [ "sanctions-record", "verification-duty", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-state-bar-generative-ai-practical-guidance", "url": "https://safelegalai.com/regulation/us-state/ca-state-bar-generative-ai-practical-guidance", "title": "Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law", "body": "State Bar of California (Committee on Professional Responsibility and Conduct)", "bodyShort": "State Bar of California", "jurisdiction": "us-state", "region": "California", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "verification", "confidentiality", "disclosure", "competence", "supervision" ], "effectiveDate": "2023-11-16T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-11-16T00:00:00.000Z", "url": "https://www.calbar.ca.gov/legal-professionals/ethics-compliance-practice-resources/ethics/ethics-technology-resources/artificial-intelligence" } ], "summary": "Practical guidance approved by the State Bar of California on 16 November 2023, the first state bar guidance on generative AI in the United States. California lawyers must not enter confidential client information into tools without adequate protections, must review and verify all output, should consider disclosing AI use to clients, and may not bill for time not spent.", "keyProvisions": [ "Confidential information may not be input into a tool lacking adequate confidentiality and security protections.", "Lawyers must critically review AI output for accuracy and bias before use.", "Disclosure to clients of generative AI use is recommended where material to the engagement.", "Fees must reflect time actually spent; AI efficiency savings cannot be billed as hours.", "Supervising lawyers must set policies for AI use by subordinate lawyers and staff." ], "relatedIncidents": [ "noland-v-land-of-the-free" ], "sources": [ { "label": "State Bar of California — AI ethics resources page", "url": "https://www.calbar.ca.gov/legal-professionals/ethics-compliance-practice-resources/ethics/ethics-technology-resources/artificial-intelligence" } ], "categories": [], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "lsbc-professional-responsibility-generative-ai", "url": "https://safelegalai.com/regulation/canada/lsbc-professional-responsibility-generative-ai", "title": "Guidance on Professional Responsibility and Generative AI", "body": "Law Society of British Columbia", "bodyShort": "LSBC", "jurisdiction": "canada", "region": "British Columbia", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "effectiveDate": "2023-11-01T00:00:00.000Z", "versions": [ { "version": "November 2023", "date": "2023-11-01T00:00:00.000Z", "url": "https://www.lawsociety.bc.ca/getContentAsset/91264ae3-9fa0-4063-88fb-69977a293f62/dfc3d011-8f63-43f6-9ed8-4b444333a1d0/Professional-responsibility-and-AI.pdf?language=en-CA", "archiveUrl": "https://web.archive.org/web/20260801052350/https://www.lawsociety.bc.ca/getContentAsset/91264ae3-9fa0-4063-88fb-69977a293f62/dfc3d011-8f63-43f6-9ed8-4b444333a1d0/Professional-responsibility-and-AI.pdf?language=en-CA" } ], "summary": "Guidance from the Law Society of British Columbia, issued November 2023, on how the Code of Professional Conduct applies when BC lawyers use generative AI. It is permissive with cautions, stressing that lawyers must verify output, keep client information confidential, remain competent in the technology, and understand that AI does not change their professional obligations.", "keyProvisions": [ "Lawyers remain responsible for the accuracy of AI-assisted work, including citations.", "Confidential client information should not be entered into public generative AI tools.", "Competence includes understanding the limits and risks of the tools used.", "Lawyers should consider whether to inform clients of AI use in their matter." ], "relatedIncidents": [ "zhang-v-chen" ], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Law Society of BC — Professional responsibility and AI (PDF)", "url": "https://www.lawsociety.bc.ca/getContentAsset/91264ae3-9fa0-4063-88fb-69977a293f62/dfc3d011-8f63-43f6-9ed8-4b444333a1d0/Professional-responsibility-and-AI.pdf?language=en-CA" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "sra-risk-outlook-ai-legal-market", "url": "https://safelegalai.com/regulation/uk/sra-risk-outlook-ai-legal-market", "title": "Risk Outlook report: The use of artificial intelligence in the legal market", "body": "Solicitors Regulation Authority", "bodyShort": "SRA", "jurisdiction": "uk", "type": "report", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality" ], "effectiveDate": "2023-11-01T00:00:00.000Z", "versions": [ { "version": "November 2023", "date": "2023-11-01T00:00:00.000Z", "url": "https://www.sra.org.uk/sra/research-publications/risk-outlook-report-use-ai-legal-market/" } ], "summary": "The SRA's November 2023 Risk Outlook report analyses how solicitors in England and Wales are adopting AI and the risks that follow. It is a risk analysis rather than a rule, taking a permissive stance while warning about hallucinated authorities, confidentiality of client data in public tools, and the need for firm-level controls.", "keyProvisions": [ "Identifies inaccurate or fabricated output as a principal risk of generative AI in legal work.", "Warns that entering client information into public AI tools may breach confidentiality duties.", "Encourages firms to understand the tools they use and to set internal policies before adoption.", "Frames AI as an opportunity for access to justice provided existing professional standards are met." ], "relatedIncidents": [ "harber-v-hmrc" ], "article": "uk-legal-profession-ai-guidance-history", "sources": [ { "label": "SRA — Risk Outlook report on AI in the legal market", "url": "https://www.sra.org.uk/sra/research-publications/risk-outlook-report-use-ai-legal-market/" } ], "country": "GB", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "ca-ns-provincial-court-ai-directive", "url": "https://safelegalai.com/regulation/canada/ca-ns-provincial-court-ai-directive", "title": "Use of Artificial Intelligence (AI) and Protecting the Integrity of Court Submissions in Provincial Court", "body": "Nova Scotia Provincial Court", "bodyShort": "NS Provincial Court", "jurisdiction": "canada", "region": "Nova Scotia", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure", "verification" ], "versions": [ { "version": "Original", "date": "2023-10-27T00:00:00.000Z", "url": "https://www.courts.ns.ca/sites/default/files/notices/Oct%202023/NSPC_Artificial_Intelligence_Oct_27_2023.pdf", "changes": "Recorded from the court's notice listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A directive of the Nova Scotia Provincial Court issued on 27 October 2023, the province's first court instrument on artificial intelligence. It requires parties to disclose the use of AI in preparing court submissions and to protect the integrity of those submissions by verifying AI-derived material before filing.", "keyProvisions": [ "Use of AI in preparing court submissions must be disclosed to the court.", "Applies to counsel and self-represented parties." ], "relatedIncidents": [], "sources": [ { "label": "Nova Scotia Courts — Provincial Court AI directive (PDF)", "url": "https://www.courts.ns.ca/sites/default/files/notices/Oct%202023/NSPC_Artificial_Intelligence_Oct_27_2023.pdf" } ], "country": "CA", "categories": [ "disclosure-filings", "verification-duty", "litigant-in-person" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-mi-judicial-ethics-opinion-ji-155", "url": "https://safelegalai.com/regulation/us-state/us-mi-judicial-ethics-opinion-ji-155", "title": "State Bar of Michigan Judicial Ethics Opinion JI-155: judicial use of artificial intelligence", "body": "State Bar of Michigan, Standing Committee on Judicial Ethics", "bodyShort": "State Bar of Michigan", "jurisdiction": "us-state", "region": "Michigan", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "competence", "supervision", "disclosure" ], "versions": [ { "version": "JI-155", "date": "2023-10-27T00:00:00.000Z", "url": "https://www.michbar.org/opinions/ethics/numbered_opinions/JI-155", "changes": "Approved 27 October 2023; published 8 November 2023" } ], "summary": "One of the earliest US judicial ethics opinions on AI. JI-155 treats technology competence as an ethical duty, requires judges to assess algorithmic bias before relying on AI, to document AI use in their reasoning on the record, and to keep decision-making human, with due-process concerns where AI is opaque.", "keyProvisions": [ "Technology competence is an ethical responsibility for all attorneys in Michigan", "Judges must assess potential algorithmic bias before relying on AI tools", "Judges must document their reasoning when using AI on the record", "Judges must distinguish between AI informing a decision and AI deciding a matter", "Judges must recognise due-process obligations where AI tools lack transparency" ], "relatedIncidents": [], "sources": [ { "label": "SBM Judicial Ethics Opinion JI-155", "url": "https://www.michbar.org/opinions/ethics/numbered_opinions/JI-155" } ], "country": "US", "categories": [ "judicial-use", "competence-duty", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "us-wv-jic-advisory-opinion-2023-22", "url": "https://safelegalai.com/regulation/us-state/us-wv-jic-advisory-opinion-2023-22", "title": "JIC Advisory Opinion 2023-22 (judicial use of artificial intelligence)", "body": "West Virginia Judicial Investigation Commission", "bodyShort": "WV Judicial Investigation Commission", "jurisdiction": "us-state", "region": "West Virginia", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "judges" ], "requirements": [ "prohibition", "verification", "confidentiality", "competence" ], "versions": [ { "version": "Original", "date": "2023-10-13T00:00:00.000Z", "url": "https://www.courtswv.gov/sites/default/pubfilesmnt/2023-11/JIC%20Advisory%20Opinion%202023-22_Redacted.pdf", "changes": "Initial publication" } ], "summary": "Answering an inquiry about judges' use of AI, the West Virginia Judicial Investigation Commission applied Code of Judicial Conduct Rules 1.2, 2.1, 2.2, 2.4(B), 2.5(A) and 2.7 and concluded that a judge may use AI for research but not to decide a case's outcome, and that drafting opinions or orders with AI demands extreme caution.", "keyProvisions": [ "A judge may use AI for research purposes but may not use it to decide the outcome of a case", "The use of AI in drafting opinions or orders should be done with extreme caution because of built-in or perceived bias", "AI is analogous to a law clerk: the judge must decide the ruling, direct the tool accordingly, and review the final draft for current law and accuracy", "Responsibility for the finished product rests solely with the judge — 'AI made me do it' is no defence", "Judges must research an AI product's data handling and refrain from inputting information that may be retained or disclosed, especially in juvenile and abuse-and-neglect matters", "Plagiarism and attribution are additional concerns when drafting with AI" ], "relatedIncidents": [], "sources": [ { "label": "WV JIC Advisory Opinion 2023-22 (official PDF)", "url": "https://www.courtswv.gov/sites/default/pubfilesmnt/2023-11/JIC%20Advisory%20Opinion%202023-22_Redacted.pdf" } ], "country": "US", "categories": [ "judicial-use", "verification-duty", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-nl-supreme-court-llm-notice", "url": "https://safelegalai.com/regulation/canada/ca-nl-supreme-court-llm-notice", "title": "Notice to the Profession and General Public: Ensuring the Integrity of Court Submissions when Using Large Language Models", "body": "Supreme Court of Newfoundland and Labrador", "bodyShort": "NL Supreme Court", "jurisdiction": "canada", "region": "Newfoundland and Labrador", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification" ], "versions": [ { "version": "Original", "date": "2023-10-12T00:00:00.000Z", "url": "https://www.court.nl.ca/supreme/files/2023-10-12-Notice-to-Profession-and-General-Public-AI-Submissions.pdf", "changes": "Recorded from the court's notice listing; the PDF was not read in full, so this record is unverified." } ], "summary": "A notice issued on 12 October 2023 by the Supreme Court of Newfoundland and Labrador to the profession and the general public, requiring anyone preparing court submissions with large language models to verify all authorities and analysis against authoritative sources before those submissions are filed with the court.", "keyProvisions": [ "Authorities generated with large language models must be verified against authoritative sources before filing.", "Applies to counsel and members of the public alike." ], "relatedIncidents": [], "sources": [ { "label": "Courts of Newfoundland and Labrador — notice (PDF)", "url": "https://www.court.nl.ca/supreme/files/2023-10-12-Notice-to-Profession-and-General-Public-AI-Submissions.pdf" } ], "country": "CA", "categories": [ "verification-duty", "litigant-in-person", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "ca-ab-tri-court-llm-notice", "url": "https://safelegalai.com/regulation/canada/ca-ab-tri-court-llm-notice", "title": "Notice to the Profession and Public: Ensuring the Integrity of Court Submissions When Using Large Language Models", "body": "Court of Appeal of Alberta, Court of King's Bench of Alberta and Alberta Court of Justice", "bodyShort": "Alberta Courts", "jurisdiction": "canada", "region": "Alberta", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "verification", "supervision" ], "versions": [ { "version": "Original", "date": "2023-10-06T00:00:00.000Z", "url": "https://albertacourts.ca/docs/default-source/qb/npp/tri-court-notice-to-profession-and-public---large-language-models.pdf?sfvrsn=713d5a82_7" } ], "summary": "A joint notice issued on 6 October 2023 by all three Alberta courts. It urges caution when referencing legal authorities or analysis derived from large language models, requires reliance exclusively on authoritative sources such as official court sites, established commercial publishers or CanLII, and requires meaningful human control over AI-assisted submissions.", "keyProvisions": [ "Exercise caution when referencing legal authorities or analysis derived from large language models.", "Rely exclusively on authoritative sources: official court and legislative websites, recognised commercial publishers, or CanLII.", "AI-assisted submissions require meaningful human control and cross-checking against reliable legal databases before filing." ], "relatedIncidents": [], "sources": [ { "label": "Alberta Courts — tri-court notice (official PDF)", "url": "https://albertacourts.ca/docs/default-source/qb/npp/tri-court-notice-to-profession-and-public---large-language-models.pdf?sfvrsn=713d5a82_7" }, { "label": "Court of King's Bench of Alberta — announcement page", "url": "https://albertacourts.ca/kb/resources/announcements/notice-to-the-profession-public---use-of-ai-in-citations-submissions" } ], "country": "CA", "categories": [ "verification-duty", "litigant-in-person", "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "nz-opc-ai-information-privacy-principles", "url": "https://safelegalai.com/regulation/nz/nz-opc-ai-information-privacy-principles", "title": "Artificial Intelligence and the Information Privacy Principles", "body": "Office of the Privacy Commissioner | Te Mana Matapono Matatapu", "bodyShort": "NZ Privacy Commissioner", "jurisdiction": "nz", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "all" ], "requirements": [ "confidentiality", "consent", "competence" ], "effectiveDate": "2023-09-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-09-21T00:00:00.000Z", "url": "https://www.privacy.org.nz/resources-and-learning/a-z-topics/ai/" } ], "summary": "The New Zealand Privacy Commissioner's expectations for agencies using generative AI under the Privacy Act 2020, published 21 September 2023 with a companion page on generative AI. The Office approves or endorses no AI technology or provider and expects a privacy assessment before an agency uses an AI tool on personal information.", "keyProvisions": [ "The Information Privacy Principles apply to personal information put into, and produced by, AI tools.", "Agencies should complete a privacy assessment before using AI tools on personal information.", "The Office does not approve or endorse any AI technology or AI technology provider." ], "relatedIncidents": [], "sources": [ { "label": "OPC — Artificial Intelligence and the Information Privacy Principles", "url": "https://www.privacy.org.nz/resources-and-learning/a-z-topics/ai/" }, { "label": "OPC — Generative Artificial Intelligence", "url": "https://www.privacy.org.nz/resources-and-learning/a-z-topics/ai/generative-artificial-intelligence/" } ], "country": "NZ", "categories": [ "data-protection", "procurement-vendor", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cz-cak-stanovisko-ai-legal-services", "url": "https://safelegalai.com/regulation/other/cz-cak-stanovisko-ai-legal-services", "title": "Stanovisko k užívání umělé inteligence (AI) při poskytování právních služeb", "body": "Czech Bar Association [Česká advokátní komora, ČAK]", "bodyShort": "ČAK", "jurisdiction": "other", "type": "ethics-opinion", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "confidentiality", "consent", "verification", "prohibition", "permissive" ], "versions": [ { "version": "1.0", "date": "2023-09-20T00:00:00.000Z", "url": "https://www.cak.cz/cs/download/priloha_4_2023_09_-ai_stanovisko.pdf", "changes": "Initial opinion, six points; issued as annex 4 to the September 2023 ČAK materials, authored by Martin Maisner." } ], "summary": "A six-point non-binding opinion of the Czech Bar Association. Lawyers may use any technology, including AI, provided they meet all statutory and professional duties. AI cannot itself provide legal services; entering client data into AI tools will normally breach confidentiality unless the client gives specific prior informed consent.", "keyProvisions": [ "Point 1: the Bar does not and will not restrict lawyers' use of technology, provided confidentiality, data-protection and ethics duties are met.", "Point 2: AI in any form cannot be a provider of legal services within the meaning of § 1 of the Act on Advocacy (Act No. 85/1996 Coll.).", "Point 4: answers addressing a concrete legal problem are legal advice reserved to authorised persons; website AI that leads a potential client to believe they are receiving lawyer's advice is 'highly inappropriate'. Using AI does not affect the lawyer's own liability.", "Point 5: AI acts autonomously and unpredictably and is potentially risky as to accuracy; entering client information — even solely to train the tool — will most likely breach the statutory duty of confidentiality, for which the lawyer is always responsible.", "Point 6: prior client consent must expressly state that the client knows AI may behave unpredictably and expressly permit sharing the information, even anonymised, with unspecified similar systems." ], "relatedIncidents": [], "sources": [ { "label": "ČAK — Stanovisko k užívání umělé inteligence (AI) při poskytování právních služeb (PDF)", "url": "https://www.cak.cz/cs/download/priloha_4_2023_09_-ai_stanovisko.pdf" } ], "country": "CZ", "categories": [ "bar-guidance", "confidentiality-client-data", "client-consent-transparency", "verification-duty", "unauthorised-practice" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ae-difc-data-protection-regulation-10", "url": "https://safelegalai.com/regulation/gulf/ae-difc-data-protection-regulation-10", "title": "DIFC Data Protection Regulations, Regulation 10 — Processing Personal Data through Autonomous and Semi-autonomous Systems", "body": "Dubai International Financial Centre Authority / Commissioner of Data Protection", "bodyShort": "DIFC Commissioner of Data Protection", "jurisdiction": "gulf", "region": "Dubai International Financial Centre", "type": "regulation", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "disclosure", "record-keeping", "risk-classification", "consent" ], "versions": [ { "version": "Enacted", "date": "2023-09-01T00:00:00.000Z", "url": "https://www.difc.com/business/registrars-and-commissioners/commissioner-of-data-protection/regulation-10", "changes": "Added to the DIFC Data Protection Regulations, in force 1 September 2023." } ], "summary": "Regulation 10 of the DIFC Data Protection Regulations, in force 1 September 2023, is the first instrument in the Middle East, Africa and South Asia to govern personal data processed through autonomous and semi-autonomous systems, including generative AI. It sets behavioural boundaries for entities that deploy or operate such systems rather than regulating algorithms directly.", "keyProvisions": [ "Applies to entities that deploy or operate autonomous and semi-autonomous systems, including AI and generative machine-learning tools, to process personal data in the DIFC.", "Requires clear notices, transparency and accountability in AI design and deployment.", "Requires systems to process data for human-defined purposes, with audit and certification obligations.", "Supplements, and is enforced alongside, the DIFC Data Protection Law." ], "relatedIncidents": [], "sources": [ { "label": "DIFC Commissioner of Data Protection — Regulation 10 (official)", "url": "https://www.difc.com/business/registrars-and-commissioners/commissioner-of-data-protection/regulation-10" }, { "label": "Regulation 10 full text (official PDF)", "url": "https://assets.difc.com/v1/media/edge/images/dubaiintern0078-difcexperie96c5-production-3253/media/project/difcexperiences/difc/difcwebsite/documents/registrars-and-commissioners/guidance-and-handbooks/lawful-processing/dp_regulation_10.pdf" } ], "country": "AE", "categories": [ "data-protection", "procurement-vendor", "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "sa-sdaia-ai-ethics-principles", "url": "https://safelegalai.com/regulation/gulf/sa-sdaia-ai-ethics-principles", "title": "AI Ethics Principles", "body": "Saudi Data and Artificial Intelligence Authority", "bodyShort": "SDAIA", "jurisdiction": "gulf", "type": "policy", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "risk-classification", "supervision", "confidentiality", "permissive" ], "versions": [ { "version": "1.0", "date": "2023-09-01T00:00:00.000Z", "url": "https://sdaia.gov.sa/en/SDAIA/about/Documents/ai-principles.pdf", "changes": "First issue, September 2023. Month-only date: the source gives the month, so the 1st is recorded as the date. The PDF was not read in full, so this record is unverified." } ], "summary": "SDAIA's national AI ethics framework, first issued in September 2023, sets non-binding principles for AI across all Saudi sectors and applies a risk-based classification to AI systems. It is the Kingdom's principal horizontal AI instrument; Saudi Arabia has no AI statute. Nothing in it addresses legal practice or courts specifically.", "keyProvisions": [ "Non-binding ethical principles applying across public and private sectors.", "Risk-based classification of AI systems.", "Emphasis on transparency, secure data practices, human oversight, fairness and accountability." ], "relatedIncidents": [], "sources": [ { "label": "SDAIA — AI Ethics Principles (official PDF)", "url": "https://sdaia.gov.sa/en/SDAIA/about/Documents/ai-principles.pdf" } ], "country": "SA", "categories": [ "ai-act-horizontal", "data-protection", "court-own-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "tw-judicial-yuan-generative-ai-judgment-drafting-2023", "url": "https://safelegalai.com/regulation/other/tw-judicial-yuan-generative-ai-judgment-drafting-2023", "title": "Judicial Yuan statement on generative-AI judgment drafting for criminal cases", "body": "Judicial Yuan, Information Department", "bodyShort": "Judicial Yuan", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "prohibition", "confidentiality" ], "effectiveDate": "2023-08-27T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-08-27T00:00:00.000Z", "url": "https://www.judicial.gov.tw/tw/cp-1887-929494-8a9fb-1.html", "changes": "Press release announcing the on-premises drafting system and the intention to adopt AI guidelines." } ], "summary": "Judicial Yuan press release of 27 August 2023 describing its generative-AI system that drafts judgment text for dangerous-driving and fraud-accessory cases, with drug offences piloted from December 2023. Fact-finding, application of law and sentencing remain entirely with the judge; the model is fine-tuned and hosted on-premises to protect privacy and official secrets.", "keyProvisions": [ "Judges alone determine facts, law and sentence; the system provides draft text only.", "On-premises deployment; case data is not sent to external services.", "Limited to specified offence types." ], "relatedIncidents": [], "sources": [ { "label": "Judicial Yuan press release (official)", "url": "https://www.judicial.gov.tw/tw/cp-1887-929494-8a9fb-1.html" } ], "country": "TW", "categories": [ "court-own-use", "ai-decision-prohibition", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "jp-moj-ai-contract-review-attorney-act-72-2023", "url": "https://safelegalai.com/regulation/other/jp-moj-ai-contract-review-attorney-act-72-2023", "title": "Relationship between AI-based contract-drafting, review and management support services and Article 72 of the Attorney Act", "body": "Ministry of Justice, Judicial System Department", "bodyShort": "Japan MoJ", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "providers", "lawyers" ], "requirements": [ "permissive" ], "effectiveDate": "2023-08-01T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-08-01T00:00:00.000Z", "url": "https://www.moj.go.jp/content/001400675.pdf", "changes": "Published August 2023 (Reiwa 5); exact day not stated." } ], "summary": "Ministry of Justice guideline of August 2023 explaining when AI-assisted contract drafting, review and management services fall outside the Attorney Act Article 72 prohibition on unauthorised legal practice. A service breaches Article 72 only if provided for remuneration, concerning a legal case, and constituting legal affairs; generative-AI services are assessed on the same framework.", "keyProvisions": [ "Three cumulative elements: purpose of remuneration, a 'legal case', and 'legal affairs'; absence of any one takes the service outside Art. 72.", "Remuneration includes indirect consideration such as subscriptions or paid upsells.", "Generative-AI-based services are judged on the same framework." ], "relatedIncidents": [], "sources": [ { "label": "Ministry of Justice guideline (official PDF)", "url": "https://www.moj.go.jp/content/001400675.pdf" } ], "country": "JP", "categories": [ "unauthorised-practice" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "cn-cac-interim-measures-generative-ai-2023", "url": "https://safelegalai.com/regulation/other/cn-cac-interim-measures-generative-ai-2023", "title": "Interim Measures for the Management of Generative Artificial Intelligence Services", "body": "Cyberspace Administration of China with the NDRC, MoE, MoST, MIIT, MPS and NRTA", "bodyShort": "CAC", "jurisdiction": "other", "type": "regulation", "status": "in-force", "appliesTo": [ "providers" ], "requirements": [ "risk-classification", "record-keeping", "consent" ], "effectiveDate": "2023-08-15T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-07-10T00:00:00.000Z", "url": "https://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm", "changes": "Issued 10 July 2023; in force 15 August 2023." } ], "summary": "Horizontal regulation on public-facing generative-AI services in force since 15 August 2023. Providers must use lawfully sourced training data, obtain consent for personal data, label generated content, undergo security assessment and algorithm filing where services have public-opinion attributes, protect user inputs, and handle access, correction and deletion requests. Not specific to legal services.", "keyProvisions": [ "Security assessment and algorithm filing for services with public-opinion attributes.", "Lawful training data, respect for intellectual property, and consent for personal information.", "Labelling of generated images and video per the deep-synthesis rules.", "Protection of user input records, with user rights and complaint channels." ], "relatedIncidents": [], "sources": [ { "label": "CAC text (official)", "url": "https://www.cac.gov.cn/2023-07/13/c_1690898327029107.htm" } ], "country": "CN", "categories": [ "ai-act-horizontal", "data-protection" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "yukon-sc-general-29-use-of-ai", "url": "https://safelegalai.com/regulation/canada/yukon-sc-general-29-use-of-ai", "title": "Practice Direction GENERAL-29: Use of Artificial Intelligence Tools", "body": "Supreme Court of Yukon", "bodyShort": "Supreme Court of Yukon", "jurisdiction": "canada", "region": "Yukon", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure" ], "effectiveDate": "2023-06-26T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-06-26T00:00:00.000Z", "url": "https://www.yukoncourts.ca/sites/default/files/2023-06/GENERAL-29%20Use%20of%20AI.pdf", "archiveUrl": "https://web.archive.org/web/20260825163735/https://www.yukoncourts.ca/sites/default/files/2023-06/GENERAL-29%20Use%20of%20AI.pdf" } ], "summary": "Practice Direction GENERAL-29, issued by the Supreme Court of Yukon on 26 June 2023, requires any party relying on artificial intelligence for legal research or submissions to advise the court of the tool used and how it was used. It followed the Manitoba direction by three days and applies to lawyers and self-represented litigants.", "keyProvisions": [ "Parties must advise the court when AI was used for legal research or submissions.", "The disclosure must identify the tool and describe its use.", "Applies in all matters before the Supreme Court of Yukon." ], "relatedIncidents": [], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Yukon Courts — GENERAL-29 (PDF)", "url": "https://www.yukoncourts.ca/sites/default/files/2023-06/GENERAL-29%20Use%20of%20AI.pdf" } ], "country": "CA", "categories": [ "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "manitoba-kb-ai-court-submissions-practice-direction", "url": "https://safelegalai.com/regulation/canada/manitoba-kb-ai-court-submissions-practice-direction", "title": "Practice Direction: Use of Artificial Intelligence in Court Submissions", "body": "Court of King's Bench of Manitoba", "bodyShort": "Manitoba King's Bench", "jurisdiction": "canada", "region": "Manitoba", "type": "practice-direction", "status": "in-force", "appliesTo": [ "lawyers", "litigants", "parties" ], "requirements": [ "disclosure" ], "effectiveDate": "2023-06-23T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-06-23T00:00:00.000Z", "url": "https://www.manitobacourts.mb.ca/site/assets/files/2045/practice_direction_-_use_of_artificial_intelligence_in_court_submissions.pdf", "archiveUrl": "https://web.archive.org/web/20260825163736/https://www.manitobacourts.mb.ca/site/assets/files/2045/practice_direction_-_use_of_artificial_intelligence_in_court_submissions.pdf" } ], "summary": "A practice direction issued 23 June 2023 by the Court of King's Bench of Manitoba, one of the first court rules on AI anywhere. When artificial intelligence has been used in the preparation of materials filed with the court, the materials must indicate how AI was used. The direction applies to lawyers and self-represented parties alike.", "keyProvisions": [ "Materials filed with the court must indicate how AI was used in their preparation.", "Applies to all submissions by counsel and self-represented litigants.", "Issued in response to concerns about reliability of AI-generated legal content." ], "relatedIncidents": [], "article": "canada-law-society-ai-guidance-history", "sources": [ { "label": "Manitoba Courts — practice direction (PDF)", "url": "https://www.manitobacourts.mb.ca/site/assets/files/2045/practice_direction_-_use_of_artificial_intelligence_in_court_submissions.pdf" } ], "country": "CA", "categories": [ "disclosure-filings" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-ed-pa-baylson-ai-standing-order", "url": "https://safelegalai.com/regulation/us-federal/us-ed-pa-baylson-ai-standing-order", "title": "Standing Order Re: Artificial Intelligence in Cases Assigned to Judge Baylson", "body": "United States District Court for the Eastern District of Pennsylvania, Judge Michael M. Baylson", "bodyShort": "E.D. Pa. (Baylson)", "jurisdiction": "us-federal", "region": "Pennsylvania", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "disclosure", "certification", "verification" ], "effectiveDate": "2023-06-06T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-06-06T00:00:00.000Z", "url": "https://www.paed.uscourts.gov/judges/michael-m-baylson" } ], "summary": "A standing order issued 6 June 2023 by Judge Michael Baylson of the Eastern District of Pennsylvania. Any attorney or self-represented party who used generative AI in preparing a filing must disclose that use in a clear statement and certify that every citation to law or the record has been verified as accurate. It applies to all his cases.", "keyProvisions": [ "Filers must disclose in a clear and plain statement that AI was used to prepare any part of a filing.", "Filers must certify that each citation to law or the record has been verified as accurate.", "Applies to attorneys and pro se litigants alike." ], "relatedIncidents": [ "mata-v-avianca" ], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "E.D. Pa. — Judge Michael M. Baylson page", "url": "https://www.paed.uscourts.gov/judges/michael-m-baylson" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "jp-ppc-generative-ai-caution-2023", "url": "https://safelegalai.com/regulation/other/jp-ppc-generative-ai-caution-2023", "title": "Alert on the use of generative AI services", "body": "Personal Information Protection Commission", "bodyShort": "Japan PPC", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "confidentiality" ], "effectiveDate": "2023-06-02T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-06-02T00:00:00.000Z", "url": "https://www.ppc.go.jp/news/press/2023/230602kouhou/", "changes": "Original alert and administrative notice to OpenAI." } ], "summary": "Personal Information Protection Commission alert of 2 June 2023 cautioning businesses and public bodies on entering personal data into generative-AI prompts: use only within the stated purpose, and do not input data that may be used for model training without a legal basis. It also issued a notice to OpenAI on collection of sensitive data.", "keyProvisions": [ "Personal data entered into generative AI must stay within the notified purpose of use.", "Businesses should confirm whether inputs are used for training and avoid inputting sensitive data.", "Administrative notice issued to OpenAI regarding sensitive personal information." ], "relatedIncidents": [], "sources": [ { "label": "PPC press page (official)", "url": "https://www.ppc.go.jp/news/press/2023/230602kouhou/" } ], "country": "JP", "categories": [ "data-protection", "confidentiality-client-data" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "us-nd-ill-fuentes-standing-order-genai", "url": "https://safelegalai.com/regulation/us-federal/us-nd-ill-fuentes-standing-order-genai", "title": "Standing Order for Civil Cases (generative AI provision), Magistrate Judge Gabriel A. Fuentes", "body": "United States District Court for the Northern District of Illinois, Magistrate Judge Gabriel A. Fuentes", "bodyShort": "N.D. Ill. (Fuentes)", "jurisdiction": "us-federal", "region": "Illinois", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "disclosure", "verification" ], "effectiveDate": "2023-05-31T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-05-31T00:00:00.000Z", "url": "https://www.ilnd.uscourts.gov/" } ], "summary": "A standing order for civil cases before Magistrate Judge Gabriel Fuentes of the Northern District of Illinois, revised 31 May 2023 to address generative AI. Any party using a generative AI tool to prepare or draft a filing must disclose that use and identify the tool, and Rule 11 verification obligations apply in full to the resulting content.", "keyProvisions": [ "Parties must disclose in the filing that generative AI was used and name the tool.", "Rule 11 applies to AI-assisted filings; counsel must verify all legal authorities.", "Use of AI for legal research must itself be disclosed if it informed the filing." ], "relatedIncidents": [], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "N.D. Ill. — Magistrate Judge Gabriel A. Fuentes judge page", "url": "https://www.ilnd.uscourts.gov/" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": false }, { "id": "us-nd-tex-starr-ai-certification", "url": "https://safelegalai.com/regulation/us-federal/us-nd-tex-starr-ai-certification", "title": "Mandatory Certification Regarding Generative Artificial Intelligence (Judge Brantley Starr)", "body": "United States District Court for the Northern District of Texas, Judge Brantley Starr", "bodyShort": "N.D. Tex. (Starr)", "jurisdiction": "us-federal", "region": "Texas", "type": "standing-order", "status": "in-force", "appliesTo": [ "lawyers", "litigants" ], "requirements": [ "certification", "verification" ], "effectiveDate": "2023-05-30T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2023-05-30T00:00:00.000Z", "url": "https://www.txnd.uscourts.gov/judges/brantley-starr" } ], "summary": "The first judge-specific generative AI order in the United States federal courts, issued by Judge Brantley Starr of the Northern District of Texas on 30 May 2023. Every attorney and self-represented litigant must certify either that no generative AI was used or that any AI-generated content was checked for accuracy by a human using print reporters or traditional databases.", "keyProvisions": [ "Certificate required on the docket for every filer appearing before the judge.", "Filers must attest that no portion of a filing was drafted by generative AI, or that all AI output was human-verified.", "Verification must use print reporters or traditional legal databases.", "Filings without the certificate may be struck; false certification exposes the filer to sanctions." ], "relatedIncidents": [ "mata-v-avianca" ], "article": "us-federal-court-ai-standing-orders", "sources": [ { "label": "N.D. Tex. — Judge Brantley Starr judge-specific requirements", "url": "https://www.txnd.uscourts.gov/judges/brantley-starr" } ], "country": "US", "categories": [ "disclosure-filings", "verification-duty" ], "lastVerified": "2026-09-03T00:00:00.000Z", "verified": true }, { "id": "eg-charter-responsible-ai", "url": "https://safelegalai.com/regulation/other/eg-charter-responsible-ai", "title": "Egyptian Charter for Responsible AI, 2023 v1.0", "body": "National Council for Artificial Intelligence, Ministry of Communications and Information Technology (MCIT)", "bodyShort": "Egypt NCAI / MCIT", "jurisdiction": "other", "type": "guidance", "status": "in-force", "appliesTo": [ "all", "providers" ], "requirements": [ "competence", "supervision", "permissive" ], "versions": [ { "version": "1.0", "date": "2023-04-19T00:00:00.000Z", "url": "https://ai.gov.eg/SynchedFiles/en/Resources/Egyptian%20Charter%20For%20Responsible%20AI.pdf", "changes": "First edition, adopted by the National Council for Artificial Intelligence." } ], "summary": "Egypt's first responsible-AI instrument, adopted by the National Council for Artificial Intelligence in April 2023. It is voluntary and sets principles of human-centred design, transparency, justice, accountability and security for AI development and deployment. It contains no provision addressed to courts, judges, lawyers or legal services.", "keyProvisions": [ "Voluntary principles-based charter: human-centred design, transparency, justice, accountability and security.", "States that accountability must always lie with natural or legal persons and that AI systems should not be given legal personality.", "Positions Egypt as an early adherent to the OECD AI Principles and the UNESCO Recommendation on the Ethics of AI, adapting them to local context.", "No sector-specific provisions for the administration of justice or the legal profession." ], "relatedIncidents": [], "sources": [ { "label": "Official PDF (ai.gov.eg)", "url": "https://ai.gov.eg/SynchedFiles/en/Resources/Egyptian%20Charter%20For%20Responsible%20AI.pdf" }, { "label": "MCIT press release", "url": "https://mcit.gov.eg/en/Media_Center/Press_Room/Press_Releases/66939" } ], "country": "EG", "categories": [ "ai-act-horizontal" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "ca-nl-law-society-ai-in-your-practice", "url": "https://safelegalai.com/regulation/canada/ca-nl-law-society-ai-in-your-practice", "title": "Artificial Intelligence in Your Practice", "body": "Law Society of Newfoundland and Labrador", "bodyShort": "LSNL", "jurisdiction": "canada", "region": "Newfoundland and Labrador", "type": "guidance", "status": "in-force", "appliesTo": [ "lawyers", "firms" ], "requirements": [ "permissive", "verification", "confidentiality", "competence" ], "versions": [ { "version": "February 2023", "date": "2023-02-01T00:00:00.000Z", "url": "https://lsnl.ca/artificial-intelligence-in-your-practice/", "changes": "Month-only date: the source gives the month, so the 1st is recorded as the date. Recorded from the law society's web page; the full resource was not read, so this record is unverified." } ], "summary": "Guidance published by the Law Society of Newfoundland and Labrador in February 2023, among the earliest Canadian law society material on artificial intelligence, applying the professional conduct duties of competence, confidentiality and accountability to the use of AI tools in the practice of law.", "keyProvisions": [ "Competence, confidentiality and accountability duties apply to AI-assisted work.", "Lawyers should understand the limits of the tools they adopt." ], "relatedIncidents": [], "sources": [ { "label": "Law Society of Newfoundland and Labrador — Artificial Intelligence in Your Practice", "url": "https://lsnl.ca/artificial-intelligence-in-your-practice/" } ], "country": "CA", "categories": [ "bar-guidance", "verification-duty", "confidentiality-client-data", "competence-duty" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "int-unesco-global-toolkit-ai-rule-of-law-judiciary", "url": "https://safelegalai.com/regulation/international/int-unesco-global-toolkit-ai-rule-of-law-judiciary", "title": "UNESCO Global Toolkit on AI and the Rule of Law for the Judiciary", "body": "United Nations Educational, Scientific and Cultural Organization", "bodyShort": "UNESCO", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "competence" ], "versions": [ { "version": "Original", "date": "2023-01-01T00:00:00.000Z", "url": "https://unesdoc.unesco.org/ark:/48223/pf0000387331", "changes": "Published 2023 (exact day not confirmed); later translated, including into Arabic." } ], "summary": "A 2023 UNESCO capacity-building resource for judicial operators explaining how AI systems work, the risks they pose to due process and human rights, and how courts can assess them. It accompanies UNESCO's judges' MOOC on AI and the Rule of Law, which has reached thousands of judicial operators, and preceded the 2025 Guidelines.", "keyProvisions": [ "Introduces AI concepts and their impact on the rule of law for judges, prosecutors and lawyers.", "Offers frameworks for assessing AI risks to fair trial, transparency and accountability.", "Companion to the UNESCO MOOC on AI and the Rule of Law, run with Oxford Said Business School and the National Judicial College." ], "relatedIncidents": [], "sources": [ { "label": "UNESCO — AI and the Rule of Law hub", "url": "https://www.unesco.org/en/artificial-intelligence/rule-law" }, { "label": "UNESDOC record", "url": "https://unesdoc.unesco.org/ark:/48223/pf0000387331" } ], "country": "INT", "categories": [ "training-cle", "judicial-use" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "cn-spc-opinions-ai-judicial-application-2022", "url": "https://safelegalai.com/regulation/other/cn-spc-opinions-ai-judicial-application-2022", "title": "Opinions on Regulating and Strengthening the Applications of Artificial Intelligence in the Judicial Field", "body": "Supreme People's Court of China", "bodyShort": "SPC", "jurisdiction": "other", "type": "policy", "status": "in-force", "appliesTo": [ "judges", "court-staff", "providers" ], "requirements": [ "prohibition", "consent", "risk-classification" ], "effectiveDate": "2022-12-08T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2022-12-08T00:00:00.000Z", "url": "https://english.court.gov.cn/2022-12/12/c_838810.htm", "changes": "Issued 8 December 2022; the official Chinese page on court.gov.cn returned 404 on 4 September 2026." } ], "summary": "Supreme People's Court normative opinion of 8 December 2022 governing AI across Chinese courts. AI output may only be an auxiliary reference: it shall never substitute for the judge's decision in any case, users may decline AI assistance, and accountability remains with the decision-maker. Systems must be lawful, secure, fair and transparent.", "keyProvisions": [ "Auxiliary-adjudication principle: AI shall not substitute for judges; results are supplementary references only and users may opt out.", "Security and legality: no harm to state secrets, network security, data or privacy.", "Transparency and credibility: systems examined by authoritative entities, with capabilities and limits disclosed.", "Fairness: prevent discrimination and remain accessible to diverse users.", "Targets of integrated AI support across judicial work by 2025 and a full framework by 2030." ], "relatedIncidents": [], "sources": [ { "label": "SPC English summary (official)", "url": "https://english.court.gov.cn/2022-12/12/c_838810.htm" }, { "label": "China Justice Observer bilingual text", "url": "https://www.chinajusticeobserver.com/law/x/the-supreme-people-s-court-the-opinions-on-regulating-and-strengthening-the-applications-of-artificial-intelligence-in-the-judicial-field-20221208" } ], "country": "CN", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true }, { "id": "hk-pcpd-guidance-ethical-development-use-ai", "url": "https://safelegalai.com/regulation/hong-kong/hk-pcpd-guidance-ethical-development-use-ai", "title": "Guidance on the Ethical Development and Use of Artificial Intelligence", "body": "Office of the Privacy Commissioner for Personal Data, Hong Kong", "bodyShort": "PCPD", "jurisdiction": "hong-kong", "type": "guidance", "status": "in-force", "appliesTo": [ "firms", "providers", "all" ], "requirements": [ "risk-classification", "confidentiality" ], "effectiveDate": "2021-08-18T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2021-08-18T00:00:00.000Z", "url": "https://www.pcpd.org.hk/english/news_events/media_statements/press_20240611.html", "changes": "The 2021-08-18 date was not independently confirmed in this pass; the document is referenced only through the PCPD's June 2024 media statement as its 2021 AI guidance and should be re-sourced to its own URL before it is relied on." } ], "summary": "The Hong Kong Privacy Commissioner's 2021 guidance setting out three Data Stewardship Values and seven Ethical Principles for artificial intelligence. It is the conceptual base on which the PCPD's 2024 Model Personal Data Protection Framework is built, and it remains the office's general statement of ethical expectations for organisations developing or deploying AI.", "keyProvisions": [ "Establishes three Data Stewardship Values and seven Ethical Principles for AI.", "Forms the conceptual base of the PCPD's 2024 Model Personal Data Protection Framework." ], "relatedIncidents": [], "sources": [ { "label": "PCPD — media statement referencing the 2021 guidance", "url": "https://www.pcpd.org.hk/english/news_events/media_statements/press_20240611.html" } ], "country": "HK", "categories": [ "data-protection", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "br-cnj-resolucao-332-2020", "url": "https://safelegalai.com/regulation/brazil/br-cnj-resolucao-332-2020", "title": "Resolucao CNJ no 332/2020: Ethics, Transparency and Governance in the Production and Use of AI in the Judiciary", "body": "National Council of Justice (Conselho Nacional de Justica)", "bodyShort": "CNJ", "jurisdiction": "brazil", "type": "regulation", "status": "superseded", "appliesTo": [ "judges", "court-staff" ], "requirements": [ "supervision", "record-keeping", "risk-classification" ], "effectiveDate": "2020-08-21T00:00:00.000Z", "versions": [ { "version": "Original", "date": "2020-08-21T00:00:00.000Z", "url": "https://atos.cnj.jus.br/atos/detalhar/3429", "changes": "First CNJ framework for AI in courts; revoked by Resolucao 615/2025, in force 14 July 2025." } ], "summary": "Brazil's first judicial AI framework, issued 21 August 2020, set principles of transparency, non-discrimination, human supervision and accountability for AI models developed or used by courts, required registration on the Sinapses platform, and barred AI-based criminal sentencing recommendations. It was replaced by Resolution 615/2025, which took effect on 14 July 2025.", "keyProvisions": [ "Principles: benefit to users, non-discrimination, publicity, transparency, user control, security and governance.", "AI models to be registered and shared through the CNJ Sinapses platform.", "Human supervision of AI-supported decisions; no AI in criminal sentencing recommendations.", "Revoked by Resolucao CNJ 615/2025." ], "relatedIncidents": [], "sources": [ { "label": "CNJ — Resolucao 332/2020", "url": "https://atos.cnj.jus.br/atos/detalhar/3429" } ], "country": "BR", "categories": [ "judicial-use", "ai-decision-prohibition", "court-own-use", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": false }, { "id": "int-cepej-european-ethical-charter-ai-judicial-systems", "url": "https://safelegalai.com/regulation/international/int-cepej-european-ethical-charter-ai-judicial-systems", "title": "European Ethical Charter on the use of artificial intelligence in judicial systems and their environment", "body": "European Commission for the Efficiency of Justice, Council of Europe", "bodyShort": "CEPEJ", "jurisdiction": "international", "type": "guidance", "status": "in-force", "appliesTo": [ "judges", "court-staff", "providers", "all" ], "requirements": [ "permissive", "risk-classification" ], "effectiveDate": "2018-12-04T00:00:00.000Z", "versions": [ { "version": "Adopted at the 31st plenary", "date": "2018-12-04T00:00:00.000Z", "url": "https://rm.coe.int/ethical-charter-en-for-publication-4-december-2018/16808f699c" } ], "summary": "Adopted by CEPEJ on 3-4 December 2018, the Charter sets five principles for AI in judicial systems: respect for fundamental rights, non-discrimination, quality and security, transparency, impartiality and fairness, and use under user control. It is addressed to policy makers, legislators and justice professionals and underpins CEPEJ's later assessment tool and generative-AI guidelines.", "keyProvisions": [ "Respect for fundamental rights: AI design and use must be compatible with the ECHR and data-protection law.", "Non-discrimination: prevent development or intensification of discrimination between individuals or groups.", "Quality and security: certified sources, multidisciplinary models, secure technological environment.", "Transparency, impartiality and fairness: processing methods must be accessible and understandable, with external audits authorised.", "Under user control: users must be informed actors in control of their choices; a judge may depart from AI output at any time." ], "relatedIncidents": [], "sources": [ { "label": "CEPEJ — European Ethical Charter (PDF)", "url": "https://rm.coe.int/ethical-charter-en-for-publication-4-december-2018/16808f699c" }, { "label": "CEPEJ — AI in justice systems", "url": "https://www.coe.int/en/web/cepej/artificial-intelligence-in-justice-systems" } ], "country": "INT", "categories": [ "judicial-use", "court-own-use", "ai-decision-prohibition", "procurement-vendor" ], "lastVerified": "2026-09-04T00:00:00.000Z", "verified": true } ] }