This week in legal AI

28 updates on the record, 2026-08-29 to 2026-09-05.

• Research · 2026-09-05
  AI chat logs, note-takers and privilege: what courts have compelled, protected and excluded, 2024 to 2026
  17 rulings and opinions on AI privilege, discovery and AI-generated evidence, 2024 to 2026, coded by question, plus what five note-taker vendors say they keep.
• Analysis · 2026-09-05
  AI in court filings: who must abstain, declare, certify or verify, and who stays silent, in 130 jurisdictions
  46 binding court instruments and 130 country records on one schema: 6 countries bind filers to disclose or certify, 10 to verify, most say nothing.
• Research · 2026-09-05
  The sanctions ledger: what courts ordered over AI-fabricated authority, 2023 to 2026, priced
  150 tracker rows, 45 with a stated figure: USD 371,542 across 34 US orders, median USD 5,000; 48 regulator referrals, 5 verified dispositions.
• Research · 2026-09-05
  Consumer AI legal services and self-represented litigants: the services, the court record and the regulators
  12 consumer legal-AI services on one schema, 41 self-represented rows in the incident tracker, two docket studies of AI pro se filings, and the regulators.
• Research · 2026-09-05
  EU AI Act obligations for law firms and legal-AI vendors, by date, article and enforcer, after the Omnibus
  Every AI Act duty that reaches a law firm or legal-AI vendor, dated after Regulation 2026/1744: 5 in force, 6 deferred to 2027, 27 national enforcers mapped.
• Research · 2026-09-05
  The bench's record: how courts and adjudicators use AI, what binds them, and where it went wrong
  106 instruments on judges' AI use and 11 decisions where an adjudicator relied on fabricated authority: 10 countries bind judges, 22 guide, 6 are drafting.
• Research · 2026-09-05
  Legal AI in India: the second-largest search market, nine court decisions and a rulebook still in draft
  India searches 'legal ai' 8,100 times a month against 9,900 in the US; its courts have acted on eight matters built on fake authority; its rulebook is in draft.
• Research · 2026-09-05
  Who searches for legal AI, what they ask, and who answers: the search-demand record, September 2026
  Ten Semrush keyword sets, 998 rows, US and UK: brand and vendor-comparison queries dominate; incident, error-rate and dated-regulation queries go unserved.
• Research · 2026-09-05
  After the referral: what bars, law societies and the SRA have done about AI misuse, 2024 to 2026
  48 matters before regulators in 7 countries: 5 dispositions verified from the regulator's own document (7 rows), 4 reported, 36 open; the fastest took 104 days.
• Research · 2026-09-04
  AI in legal practice: what 130 country records require at September 2026
  A dataset-first report on AI rules for lawyers and courts across 130 records, with provisional countries separated from checked counts.
• News · 2026-09-04
  D.C. Court of Appeals strikes Deutsche Bank brief over Google AI citations
  The D.C. Court of Appeals struck a Deutsche Bank appellee brief after counsel used Google's generative AI search tool without verifying four nonexistent cases.
• News · 2026-09-04
  India's Supreme Court sets aside customs penalty over AI-hallucinated citations
  The Supreme Court of India set aside a Rs 425.28 crore customs penalty after finding fake or misused AI-generated authorities in the order.
• News · 2026-09-04
  Indonesia's Supreme Court bars uploading draft judgment templates to AI services
  A 2 September 2026 Mahkamah Agung letter circulates draft criminal-code judgment templates and bars uploading the drafts to AI services.
• Analysis · 2026-09-04
  Ellis George and K&L Gates: the $31,100 AI-research sanction, and the re-filing that made it worse
  A C.D. Cal. Special Master sanctioned Ellis George and K&L Gates $31,100 after an AI-drafted brief cited fake cases and was re-filed without disclosure.
• News · 2026-09-04
  Virginia appeal brief drew CLE sanction after 25 phantom quotes
  The Virginia Court of Appeals ordered counsel to take AI or appellate-practice CLE after an appellee brief attributed 25 phantom quotations to Virginia cases.
• Analysis · 2026-09-04
  May judges and court staff use AI in their own work?
  The regulation map records 39 country answers on judicial AI use: binding rules, guidance, proposals and one case-law answer.
• Analysis · 2026-09-04
  Must lawyers verify AI output before relying on it?
  Across 130 country records, 50 have a binding, guidance, proposed or case-law answer on verification of AI output.
• Analysis · 2026-09-04
  Sullivan & Cromwell's AI hallucinations: what the firm told the bankruptcy court, and what happened next
  Sullivan & Cromwell disclosed AI-hallucinated citations in a Chapter 15 emergency motion in April 2026. The letter, the 28-citation count, and the docket since.
• Incident · United States (state) · 2026-09-03
  Douglas v. Deutsche Bank National Trust Co., No. 24-CV-1099 (D.C. Ct. App.)
  In the District of Columbia Court of Appeals, Deutsche Bank's appellee brief cited four nonexistent authorities after counsel used Google's generative AI search tool without verification. The court struck the brief, referred the matter to the Office of Disciplinary Counsel and sent the sanctions-authority question to its Rules Committee.
  Primary source: Douglas v. Deutsche Bank National Trust Co., No. 24-CV-1099 (D.C. Ct. App. 3 Sep 2026), published order PDF via CourtListener storage
• Analysis · 2026-09-03
  EU AI Act update — September 2026: the Digital Omnibus is law and Article 50 applies
  EU AI Act, July to August 2026: Regulation (EU) 2026/1744 in force, high-risk duties deferred to 2027 and 2028, Article 50 and GPAI enforcement live.
• Research · 2026-09-03
  Legal AI incident report, Q3 2026 to 3 September: 54 incidents, 33 in US courts, 20 regulator outcomes
  1 July to 3 September 2026: 54 tracker incidents in 11 jurisdictions, USD 132,888 in stated US penalties, 20 rows with a regulator referral or disposition.
• Analysis · 2026-09-03
  Summer 2026 in legal AI regulation: SRA warning notice, mandatory Irish declaration and a deferred EU Act
  Summer 2026 reviewed: the SRA's AI warning notice, Ireland's mandatory AI declaration in civil filings and the EU's deferred high-risk AI Act duties.
• Weekly briefing · 2026-09-03
  This week in legal AI safety: tracker reaches 122 incidents; SRA AI warning notice; Ireland's HC 142 live
  Issue 1: 50 incidents added, the SRA's misuse-of-AI warning notice, Ireland's Practice Direction HC 142, deferred EU high-risk duties, 128 tools on record.
• Incident · India · 2026-09-02
  Vijay Ghanshyam Gadiya v. Union of India, 2026 INSC 947
  Extending Pooja Ramesh Singh from tribunals to executive adjudicators, the Supreme Court of India set aside a customs penalty order of INR 4.25 billion, and the High Court judgment upholding it, because the order relied on AI-hallucinated precedents. Justice Datta held even an iota of fake material vitiates a decision and left disciplinary action to the officer's appointing authority.
  Primary source: Supreme Court of India — Vijay Ghanshyam Gadiya v. Union of India, 2026 INSC 947 (official PDF)
• Rule · Mahkamah Agung · 2026-09-02
  Letter No. 112/WKMA.Y/HK2.1/IX/2026: draft judgment-writing guidance and templates for the new Criminal Codes
  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.
  Primary source: Mahkamah Agung — Original
• Rule · Supreme Court of India · 2026-09-02
  Vijay Ghanshyam Gadiya v Union of India, 2026 INSC 947
  Original judgment (Datta and Nagu JJ). 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.
  Primary source: Supreme Court of India — Judgment
• Rule · Lietuvos advokatūra · 2026-09-01
  AI in the advocate's practice: opportunities, tools and safe use [DI advokato praktikoje: galimybės, įrankiai ir saugus naudojimas]
  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. 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.
  Primary source: Lietuvos advokatūra — 1.0
• Incident · United States (state) · 2026-08-31
  State v. Presti, No. CAAP-24-0000826 (Haw. Ct. App.)
  On a Hawaii criminal appeal, self-represented appellant Cody Presti filed an opening brief citing 95 cases, 11 of which the court could not locate; six citations pointed to unrelated decisions. The Intermediate Court of Appeals treated arguments supported by nonexistent authority as waived, affirmed the conviction and warned that similar filings can trigger sanctions.
  Primary source: State v. Presti, No. CAAP-24-0000826 (Haw. Ct. App. 31 Aug 2026), published opinion (official PDF)

Source: SafeLegalAI (safelegalai.com), published by Cognesio LLP. Data CC BY 4.0. As of 2026-09-05. Records: safelegalai.com/ · Feed: RSS