Regulation Tracker · Country · International bodies
International bodies
Extensiven/aupdated 2026-09-04verified 2026-09-04
As of 4 September 2026 no international body imposes binding rules on AI in legal practice; the Council of Europe Framework Convention on AI is signed but not in force. The field is soft law: CEPEJ's Ethical Charter and generative-AI guidelines, UNESCO's court guidelines, CCBE and IBA practitioner guidance, and arbitration guidance from SVAMC, Ciarb, VIAC and SCC.
As of 2026-09-04, International bodies has 0 binding rules, official guidance in 16 categories, 1 proposal and 0 positions set only by case law across the 20 categories in the map. In 3 categories we found nothing; in 0 the position is not yet determined.
How to read this page
Binding ruleBinding rule means a rule in force from a court, statute, regulator or equivalent authority.
Official guidanceOfficial guidance means an official non-binding position from a judiciary, regulator, bar or public body.
Proposed / consultationProposed / consultation means a public bill, draft rule or consultation that is not yet in force.
Set by case lawSet by case law means no rule was found, but courts have set the position in decisions.
Nothing foundNothing found means the category was checked and no rule, guidance, proposal or case-law position was found.
Not yet determinedNot yet determined means research did not establish a reliable position yet.
No international court requires AI certification in filings. Arbitration soft law addresses disclosure: SVAMC Guideline 3 makes disclosure case-by-case; the Ciarb Guideline art. 7 lets tribunals impose continuing disclosure duties; CEPEJ(2025)18 requires transparency where judicial generative-AI use affects proceedings.
A consistent soft-law duty to check output: CEPEJ(2025)18 requires outputs to be verified before reliance; the CCBE Guide of 2 October 2025 makes lawyers responsible for checking invented citations; SVAMC Guideline 4 and Ciarb 8.3 require parties and arbitrators to verify AI output; AAA-ICDR tells arbitrators to cross-reference primary sources.
Two global reference texts: the CEPEJ Guidelines on generative AI for courts, adopted 5 December 2025 for 46 Council of Europe states, and the UNESCO Guidelines for the use of AI systems in courts and tribunals of 3 December 2025. Both sit on the 2018 CEPEJ Ethical Charter and CCJE Opinion No. 26.
Non-binding but uniform: CEPEJ(2025)18 states judicial power is exclusively for courts and generative-AI output is never binding; UNESCO principles 1.13-1.15 say AI shall not replace judicial decision-making; SVAMC Guideline 6, Ciarb 8.2, VIAC para 2 and the SCC Guide bar arbitrators from delegating decisions.
Arbitration soft law only: SVAMC Guideline 5 asks parties to identify AI-generated material and its limits; Ciarb art. 6 covers admissibility, bias and source-data linkage; VIAC para 6 leaves admissibility and weight to the tribunal; the SCC Guide urges marking of AI-generated content. No international-court evidence rule found.
The CCBE Guide ties professional secrecy to what is entered into tools and where data is processed; CEPEJ(2025)18 bars confidential case data in unsecured tools; SVAMC Guideline 2, Ciarb 2.2, VIAC para 3 and AAA-ICDR apply the same rule to arbitration, barring party names and case specifics without data-protection guarantees.
The CCBE Guide requires lawyers to understand tool limitations and flags the EU AI Act article 4 literacy duty; the IBA and CAIDP report recommends AI-literacy programmes; SVAMC Guideline 1, VIAC 4.1 and AAA-ICDR extend competence expectations to arbitration participants. No binding rule exists.
Only indirect guidance: the CCBE recommends firm-level AI policies; the IBA and CAIDP report urges ethics rules covering proper supervision and use of AI tools; Ciarb 8.3 tells arbitrators to keep AI output under appropriate supervision. No supervisory duty is framed as a rule.
The IBA and CAIDP report of September 2024 recommends that ethics guidelines include disclosure obligations regarding the use of AI, and the CCBE Guide addresses transparency toward clients as good practice. No international body requires client consent; national bars decide.
The CCBE Guide on generative AI for lawyers of 2 October 2025 is the reference for European bars; the IBA and CAIDP report The Future is Now and the IBA ANLBS guide on AI as a professional tool are the global bar-level texts. All are non-binding.
The CEPEJ Ethical Charter of 4 December 2018 and its Assessment Tool, revised 5 June 2025, govern deployment of AI in courts; CEPEJ(2025)18 requires pseudonymised data and public control of infrastructure; the 2025 UNESCO Guidelines add lifecycle monitoring and the capacity to disable systems.
No legal-AI-specific data-protection guidance exists at international level. CEPEJ(2025)18 requires pseudonymisation and judiciary-controlled infrastructure; the ICC Note to Parties of 1 June 2026 reminds tribunals that the GDPR and EU AI Act may apply. CETS 225 article 11 on privacy is not yet in force.
No international tribunal has sanctioned AI misuse. Ciarb 7.6-7.7 lets tribunals draw adverse inferences and reflect undisclosed AI use in costs. The only reported challenge is domestic to the United States: LaPaglia v Valve, filed 8 April 2025, alleging an arbitrator outsourced the award to AI; the outcome is unverified.
Dense soft law: SVAMC Guidelines of April 2024, the SCC Guide of October 2024, the Ciarb Guideline of March 2025 updated September 2025, the VIAC Note of April 2025, AAA-ICDR arbitrator guidance and the JAMS AI Disputes Rules. The ICC, LCIA, SIAC and HKIAC have no dedicated guidance.
Recommended but never mandatory: the 2025 UNESCO Guidelines tell judicial organisations to run AI training programmes; CCJE Opinion No. 26 says judges must be kept up to date on technology; the IBA and CAIDP report recommends AI-literacy programmes. UNESCO also runs a judges' MOOC.
The CEPEJ Assessment Tool gives court decision-makers criteria for evaluating AI tools against the Ethical Charter; the 2025 UNESCO Guidelines recommend procurement standards with risk and rights-impact review; the SCC and VIAC tell participants to do due diligence; AAA-ICDR asks arbitrators to select tools with robust data security.
The Council of Europe Framework Convention on AI opened for signature in Vilnius on 5 September 2024; the Treaty Office chart of 3 September 2026 shows 20 signatures without ratification and one ratification by the EU, so it is not in force. The OECD AI Recommendation is non-binding.
No international body has issued guidance for self-represented litigants using AI as of 4 September 2026. UNESCO principles on equality of arms and Ciarb 6.3 touch on unequal AI access but are addressed to courts and tribunals, not to litigants.
No international guidance on billing for AI-assisted work found as of 4 September 2026. The IBA and CAIDP report only recommends fora on moving from hourly to fixed or value-based fees; it sets no billing rule.
No rule, guidance or reported case found at international level as of 4 September 2026; unauthorised-practice regulation is national. The IBA ANLBS guide only surveys national positions.
Court incidents on the record in International bodies
No incidents in the tracker for this country yet.
Cite this page
SafeLegalAI Legal AI Regulation Tracker, "International bodies" (updated 2026-09-04, verified 2026-09-04), https://safelegalai.com/regulation/country/int (accessed 2026-09-04). Data: CC BY 4.0.
SafeLegalAI is a research publication by Cognesio LLP, not a law firm. Nothing here is legal advice, and no lawyer–client relationship arises from reading it. Rules change; always check the official document linked on each record and take advice on your own situation. Researched and drafted with AI assistance; verified against primary sources and edited by Cognesio LLP. The linked official documents are the record — our summaries are not the law and are not legal advice.
International bodies