AI Law Wiki News for November 2026

AI law and policy news first covered in November 2026. Each story links to its sources and relevant wiki coverage where available.

November 3, 2026

• New York State Bar adopts AI ethics guidelines for lawyers — The New York State Bar Association’s House of Delegates adopted AI ethics guidelines on October 31. The association advises lawyers to verify outputs, supervise vendors and agentic tools, address bias and conflicts, and disclose relevant AI use. Its billing guidance distinguishes time spent checking outputs from time spent learning a tool and treats ordinary acquisition costs as overhead. These are bar-association guidelines, not newly enacted court rules. NYSBA — complete adoption announcement
• Georgia judge orders litigant to certify accuracy of pleadings — In an October 27 order in Williams v. Bowen, a federal judge in Georgia warned about AI hallucinations and required the self-represented plaintiff to sign a statement verifying the facts, authorities and legal theories in every pleading. The order warns that noncompliance can lead to sanctions, including dismissal. It does not find that Williams used AI or fabricated citations, and it imposes no monetary AI sanction. Court — complete October 27 order
• Chinese financial regulator warns about concentrated AI investment — China’s financial-regulator deputy Xiao Yuanqi warned on November 3 that concentrated AI investment could increase leverage and crowd out traditional industries. He urged policymakers to prepare frameworks addressing economic and financial risks, Reuters reports. The remarks do not announce a new binding rule. Reuters via WIMZ — complete report
• EU court adviser says AI news answers may infringe publishers’ rights — In a nonbinding opinion in Like Company v Google Ireland on November 3, Advocate General Maciej Szpunar said real-time retrieval and reuse of news by AI chatbots can engage press publishers’ rights in specified circumstances. His analysis distinguishes unprotected facts from recognisable expression and addresses access restrictions and answers that substitute for original publications. He recommends treating the model-training questions as inadmissible. The Court of Justice has not issued its judgment. Court of Justice — complete press release
• UK plans consultation on AI-agent loss-of-control safeguards — A November 3 government factsheet outlines plans to consult on safeguards for AI-agent loss-of-control incidents affecting regulated networks and information systems. Proposed measures cover security and resilience, notifying affected organisations, preserving incident evidence and ministerial post-incident reviews. Implementation would require secondary legislation after the Cyber Security and Resilience Bill receives Royal Assent, consultation and parliamentary approval. The factsheet does not create immediate duties or a comprehensive AI governance regime. UK government — agentic AI resilience factsheet
• China outlines AI legislation and lifecycle safety standards — China’s National Development and Reform Commission outlined plans on November 3 to accelerate foundational AI legislation and develop lifecycle safety standards and ethical guidance, National Business Daily reports. The plans also address regulatory sandboxes, monitoring employment displacement and protecting people against algorithmic discrimination, deepfakes and privacy harms. These are policy and legislative intentions, rather than an enacted AI statute or newly operative duties. National Business Daily — complete report
• Indonesian legislative draft proposes safeguards for AI-assisted lawmaking — Indonesia’s parliamentary expertise agency presented a legislative-process draft on November 3 that would permit AI assistance in drafting legislation while requiring human oversight, accuracy, transparency, accountability, information security and personal-data protection, Suara reports. The presentation remains an early drafting step, before establishment of a working committee; the proposal is not enacted law. Suara — complete two-page report
• CPS cited fictional case; court rejects challenge based on error — The High Court criticised a CPS prosecutor’s reliance on a nonexistent authority in a prosecution-decision record, but said the error did not change the outcome and rejected the proposed challenge based on it, Legal Futures reports. The Guardian attributes the error to an AI-generated search summary. The judicial review was dismissed; the reported ruling does not impose an AI-specific fine. Legal Futures — complete report · The Guardian — AI-search attribution
• Brazilian audit recommends stronger safeguards for court AI drafts — Brazil’s federal audit court recommended stronger prompt-injection testing, support for human review and factual-fidelity checks for the TRT4 labour court’s Galileu drafting system. Its October 28 announcement of an October 14 decision says the tool was reliable in the central dimensions examined, while stressing that this was not unrestricted or permanent certification. Judges remain responsible for decisions. The audit examined system controls, rather than the merits of judicial rulings. TCU — complete audit announcement
• Singapore institute begins work on financial-sector AI safeguards — The Monetary Authority of Singapore announced on November 3 that the Future of Finance Institute has begun operations. Its initial AI programmes will develop practices for safe agent deployment and resources to implement MAS’ AI risk-management guidelines, including risk assessments, testing, assurance and third-party AI. Further tools and details remain planned. The announcement starts a public-private implementation initiative; it does not announce a new binding AI rule. MAS — complete commencement announcement

November 2, 2026

• WA police end facial-recognition trial pending safeguards review — Western Australia police have removed live facial-recognition vans from public deployment while evaluating a five-month trial, the Guardian reports. Police say 100 deployments scanned 1,380,882 faces and led to 122 arrests, with nine incorrect alerts; Commissioner Col Blanch said none of those false alerts led to an arrest. Civil Liberties Australia disputes how police present the error rate and raises concerns about scanning people without reasonable suspicion. The internal evaluation will examine performance, community impacts and safeguards. These are police-reported results and competing privacy concerns, not an independent accuracy assessment or a finding that the trial was unlawful. The Guardian — Ima Caldwell and Penry Buckley
• FCC notice fixes comment deadlines for AI-enabled sensing proposals — A November 2 Federal Register notice sets December 2 comments and January 4, 2027 replies in the FCC’s ultra-wideband proceeding, ET Docket 26-245. The notice asks whether AI-based sensing systems need a distinct device category and what technical, operational and interference safeguards should apply. It accompanies broader proposals to modernize ultra-wideband equipment rules. The FCC adopted the underlying proposal on September 30 and released it October 1; this is the Federal Register publication and comment timetable, not a newly adopted final AI rule. Federal Register — FCC 26-66 notice · FCC UWB and AI-sensing proposal
• England expands AI planning-assessment pilot to 15 councils — An October 30 government announcement and October 31 DeepMind report describe a 12-month expansion of PlanAssist from three English councils to 15. The tool gathers application documents, planning policies and site constraints and drafts assessments for officers to review. Officers retain responsibility for decisions, and the government says the software will not be mandatory. DeepMind reports a 58% reduction in time spent on assessment tasks in early testing; that is a reported pilot result, not independent proof that every planning decision will be faster. Wider availability in 2027 depends on the planned rollout and successful testing. UK housing ministry — complete announcement · Google DeepMind — pilot and beta details
• Scissero and Vetta seek Weil technology, know-how and lawyers — Artificial Lawyer reports that AI-enabled firms Scissero and Vetta are seeking technology, know-how and lawyers from Weil Gotshal. Scissero’s founder says it has submitted an expression of interest in assets including DealVision 360 and is discussing recruitment; Vetta is reportedly pursuing precedents, playbooks and knowledge-management assets. Scissero says client confidentiality, privilege and any required consents would need separate assessment. These are reported interests and bids, not completed acquisitions or authorization to use confidential client data. They are distinct from Crosby’s previously reported approach. Artificial Lawyer — Scissero interview · Artificial Lawyer — Vetta report and interview
• Tech association urges proportionate EU AI-scraping guidance — In an October 30 submission published on its website November 1, Chamber of Progress asks the EDPB to preserve workable legitimate-interest assessments for AI training, avoid treating detailed source inventories or public impact assessments as universal requirements, and use proportionate safeguards for incidental sensitive data. It also urges the Board to wait for relevant Digital Omnibus legislative changes before finalizing the guidelines. These are a technology-industry association’s recommendations on draft Guidelines 03/2026, not an EDPB decision or a new exemption from GDPR. Chamber of Progress — complete seven-page submission
• Musician and policy groups defend AI-training fair use in OpenAI case — An October 30 proposed amicus brief by Trevor Lawrence Jr., Chamber of Progress and Abundance Institute argues that extracting non-expressive patterns for model training is transformative and supports human creativity. It cites musicians’ use of AI for arrangements, learning and lower-budget production and asks the New York court to grant OpenAI summary judgment on the training uses addressed. Filed as ECF 2462-1 and published on the association’s website November 1, the brief expressly takes no position on how the works were obtained. These are amici’s arguments, not judicial findings or a verified order accepting the brief. Proposed amicus brief — complete ECF 2462-1 · OpenAI copyright litigation — background
• Copyright Office corrects AI-training wording in data-mining notice — A November 2 Federal Register correction adds the word “not” before “permitted by the exemption” in the Copyright Office’s October 28 DMCA notice. The affected sentence concerns circumventing access controls to compile AI-model training datasets in the discussion of the literary-work research text-and-data-mining exemption. The correction aligns the sentence with the surrounding discussion of non-exempt uses. It corrects the proposed-rule preamble; it does not enact a new exemption or finalize the intended research-exemption renewal previously reported. Copyright Office — complete correction notice · October 28 notice — affected section II.J
• Google and Anthropic disclose limits on Japan training-data inquiries — Google and Anthropic’s October 26 disclosures, linked from Japan’s official Principle Code registry updated October 30, explain limits on answering inquiries about whether particular URLs were used for AI training. Google says it lacks a mechanism to give third parties reliable URL-level confirmation and cites technical, trade-secret and security concerns. Anthropic says identifying specific training URLs or searching by a comparison URL is currently difficult and that it generally cannot answer disclosure requests under Principles 2 and 3. These are company explanations under the voluntary code, not government certification or a determination that their copyright practices are lawful. Japan — official disclosure registry · Google — complete six-page declaration · Anthropic — complete three-page declaration
• Watchdog examines disputed NHS claims about Palantir platform benefits — The Guardian reports that 29 trusts listed by NHS England as benefiting from Palantir’s federated data platform said in freedom-of-information responses that the system was not live or they were not using its applications. The Office for Statistics Regulation said it would take the benefits claim forward under its casework function. NHS England says nearly 140 trusts report benefits, including back-office analysis and data submissions; some trusts may therefore use limited functionality without operational use of the apps. Palantir defended the platform’s contribution to care. The report renews parliamentary scrutiny of the £330 million public contract; it does not establish that every listed trust received no benefit. The Guardian — full investigation
• Argentina courts AI data centres with tax incentives still awaiting approval — AP reports that Argentina’s proposed 15% corporate tax rate for qualifying AI investments awaits Senate approval. Copyright and cross-border data rules remain unsettled, while Patagonia data-centre permits are pending. Milei’s hands-off pitch is a policy position, not an enacted blanket exemption. AP — full reported investigation
• Maryland reports five complaints under election-disinformation law — The Baltimore Sun reports that Maryland election officials received five complaints under the state’s new law targeting AI-generated election disinformation and found one violation, involving an advertisement with the wrong Election Day date. Officials say the elections board’s corrective authority covers false information about voting, results and voting rights; other complaints go to the state prosecutor. The report describes early enforcement and limits of an existing law, rather than a new enactment. The Baltimore Sun via FOX45 — full report
• UK and Germany plan to expand defence cooperation to AI — Euronews reports that Britain and Germany plan joint testing, development and eventual deployment of military AI by expanding their Cyber Lighthouse cybersecurity partnership. British Defence Secretary Wes Streeting outlined the plan ahead of November 2 talks with Boris Pistorius. Government and industry representatives are to meet in spring 2027 to work out implementation. No specific AI systems or budgets have been announced; this is a cooperation plan, not a completed deployment or new autonomous-weapons rule. Euronews — full report
• Tenth Circuit rejects unsupported AI-based fee reduction — In its October 30 Flux Drive v. Sparks opinion, the Tenth Circuit affirmed dismissal of certain defendants for lack of personal jurisdiction and related fee awards. It rejected the plaintiffs’ argument that AI and online research should have reduced billed hours because they identified no evidence of excessive billing. The court acknowledged that these tools may reduce research time, while stressing lawyers’ verification duties. The ruling assesses the record in this case; it does not establish that AI use can never justify a lower fee. Tenth Circuit — complete October 30 opinion
• New York judge sanctions lawyer over unverified AI assertions — In an October 28 order in PC-59 Doe v. Jewish Board of Family and Children’s Services, Justice Sabrina Kraus imposed a $250 sanction after defence counsel acknowledged using an AI tool without verifying its output, producing a false quotation and improper legal assertions. The judge considered counsel’s candour, acceptance of responsibility and apparent first offence. The order requires payment to the Lawyer’s Fund for Client Protection and proof of compliance; it applies existing professional and sanctions duties rather than banning AI tools. New York Supreme Court — complete October 28 order
• Delaware reportedly subpoenas OpenAI over safety and governance — Spotlight Delaware reports that state justice officials issued an investigative subpoena to OpenAI in October, seeking safety and governance records after the July Hugging Face incident. The outlet cites unnamed officials and a person familiar with the demand. Delaware oversees the nonprofit controlling OpenAI; the reported inquiry concerns its safety commitments following the 2025 restructuring. OpenAI did not respond to the outlet. An investigative demand is not a finding of wrongdoing. Spotlight Delaware — full report
• Newspapers sue Brave over alleged copying for AI services — The Verge reports that The New York Times and, separately, the New York Daily News, Chicago Tribune and Orlando Sentinel sued Brave over alleged unauthorized copying for AI services and paywall bypassing. Brave did not immediately respond to the outlet. The claims have not been adjudicated. The Verge — full report
• Brazil’s customer-service decree brings human-help requirements into force — Brazil’s Decree 13.135 took effect on November 2, twenty days after publication. It permits AI-assisted customer service while requiring a human option, including at each menu level, during prescribed hours. Covered providers must generally offer human assistance from 8 a.m. to 8 p.m. daily, subject to sector-specific exceptions. The decree covers qualifying federally regulated services and specified central-bank-supervised digital financial services; it does not impose these rules on every business. Brazil — Decree 13.135, full text · Exame — commencement report · Brazil SAC decree — human assistance
• New York proposal would expand independent frontier-AI oversight — New York Senator Andrew Gounardes announced a proposed RAISE Act expansion on October 29. Bill S10731, introduced October 28 and referred to the Senate Rules Committee, would establish independent, state-licensed frontier-AI examiners, annual audits, periodic risk assessments and incident reporting. The sponsor also proposes powers to restrict computing resources or seek court-ordered deployment limits in serious cases. These are proposed safeguards, not enacted duties or an operating examination regime. New York Senate — sponsor’s full release · S10731 — official status · New York RAISE — S10731 expansion proposal
• OpenAI outlines utility agreements to protect other ratepayers — OpenAI announced a Grid Protectors program for future U.S. computing campuses, describing utility arrangements intended to limit effects on other electricity customers. The company says its planned Effingham County, Georgia campus has a Georgia Power agreement allowing up to 1,000 megawatts of flexible demand. It pledges to cover infrastructure and service costs and provide financial assurances if projects are delayed or cancelled. Arrangements elsewhere remain to be negotiated; these are company commitments and development plans, not a new legal mandate or delivered savings. OpenAI — full program announcement
• Entrepreneur–Meta case heads toward May trial and amended complaints — The Trademark Lawyer reports that an October 27 filing confirms a May 3, 2027 trial in Entrepreneur Media’s AI-copyright case against Meta, following an October 14 scheduling order. A later October 29 court minute entry permits Entrepreneur and the Three Cs plaintiffs to file amended complaints by November 6, with advance copies to Meta by November 3. The court has not resolved copyright liability. The later entry updates the report’s account that Entrepreneur was still expected to seek amendment permission. The Trademark Lawyer — full report · Court — October 29 minute entry
• FCA reports AI-related service failure and sensitive-data leaks — The FCA’s November 2 wholesale-bank review describes inaccurate AI-generated code recommendations passing human review and causing a service outage. It also reports sensitive-data leaks from unapproved AI tools. Bank internal audits identified incomplete AI inventories, undocumented production exceptions, misleading approval records and weaknesses in kill-switch training and testing. The regulator asks firms to assess how the findings apply to their activities. These are supervisory observations under existing requirements, not a new AI statute or a finding that every bank has these weaknesses. FCA — complete supervisory review
• Hong Kong SFC plans guidance for more autonomous AI — Hong Kong’s SFC says it will develop principles-based guidance for more autonomous AI and a cybersecurity framework with stronger requirements for higher-risk firms. Its November 2 survey announcement emphasizes human accountability, defined authority and proportionate monitoring rather than manual approval for every action. Thirty percent of twenty surveyed financial groups reported using agentic AI and another quarter were piloting it. The findings are self-reported and do not represent every licensed firm; the announced guidance and cyber framework remain plans. SFC — complete survey announcement · SFC — complete November 2 remarks
• HKMA launches agentic-finance governance project — Hong Kong’s monetary authority launched Project SAGE and a cross-sector expert group on November 2 to explore AI agents in finance. The project covers use cases, verification of agents’ identities and delegated authority, interoperability, and legal questions including liability and consumer protection. A companion announcement says the authority published a discussion paper seeking industry feedback on banks’ governance of agentic AI. The initiative develops supervisory thinking and practical applications; it does not itself enact a new liability regime. HKMA — full Project SAGE announcement · HKMA — discussion-paper announcement
• Technology and music groups press opposing AI-training arguments — CCIA announced on November 2 that it co-filed an OpenAI copyright amicus brief with NetChoice and TechNet, arguing that AI training is transformative fair use and rejecting market-dilution claims. Separately, Music Business Worldwide reports that the RIAA, NMPA and seven other music organizations submitted an October 30 proposed brief supporting the plaintiffs and arguing that AI outputs compete with human work and that licensing is practical. These are advocacy positions; neither source establishes a ruling accepting their arguments or deciding fair use. CCIA — complete amicus announcement · Music Business Worldwide — full report · OpenAI copyright litigation — background
• Cassidy seeks feedback on health-AI oversight and payment draft — Senate HELP Chair Bill Cassidy released the PATIENT AI Act discussion draft on October 29. His announcement proposes clarifying FDA and health-information-technology oversight, allowing rolling review of software updates, developing a payment pathway for AI tools and directing health-department privacy and security guidance. Stakeholder feedback is requested by December 1. This is a discussion draft, not an enacted law or a verified numbered bill introduced in Congress. Senate HELP — full sponsor announcement · PATIENT AI — discussion draft
• Anthropic steers Claude Code away from recreating existing games — Anthropic has added instructions to Claude Code on Opus 5.5 to steer users away from recreating specific existing games, Ars Technica and The Verge report. Engineer Addy Osmani confirmed the change on October 30, citing the company’s policy against intellectual-property infringement. Original games, genre-based games and projects using users’ own designs and assets remain intended uses. This is a company safeguard, not a court determination that every game recreation infringes copyright or a guarantee that the controls work. Ars Technica — complete written report · The Verge — complete written report
• U.S. agencies withhold AI-review framework records in FOIA case — In an October 30 response in Protect Democracy’s FOIA lawsuit, federal agencies say they are withholding all 60 responsive pages, comprising 30 records, concerning voluntary frontier-AI review and the GOLD EAGLE clearinghouse. They invoke national-security, statutory, confidential-commercial-information and privilege exemptions. The accompanying index includes a seven-page voluntary framework and trusted-partner communications. These are the agencies’ withholding claims, not a court ruling upholding the exemptions. ONCD — complete response and withholding index · Court filing — September 15 joint status report
• Ofcom assesses facial age estimation for proposed under-16 restrictions — Ofcom’s October 29 assessment identifies facial age estimation, photo-ID matching, digital identity services and banking checks as methods capable of highly effective age assurance at 16. It cautions that facial estimates may need a second verification check, some teenagers lack suitable ID, and circumvention cannot be eliminated. The assessment informs proposed age restrictions and forthcoming guidance; it does not establish that the restrictions are already operating. The ICO separately stresses privacy and existing protections for underage users. Ofcom — complete assessment summary · Ofcom — full announcement and qualifications · ICO — complete October 29 statement
• UK cyber agency urges preparedness for AI-assisted infrastructure attacks — The NCSC’s annual review warns that AI-powered ransomware could target UK critical infrastructure by 2028, AP reports. Cyber chief Richard Horne urges organisations to invest in resilience and address older vulnerabilities. This is a threat assessment and call for action, not a finding that such an AI-driven attack has already occurred or a new legal requirement. AP — complete cyber-policy report
• Copia argues copyright should not be used to halt AI development — In an October 30 account of its OpenAI and Microsoft amicus brief, the Copia Institute argues that copyright should not prevent learning from works or lawful expression enabled by AI. It contends that using copyright to obstruct those activities would conflict with constitutional limits and the right to receive information. Copia acknowledges AI-related harms but argues they require better-targeted responses. These are the advocate’s arguments, not a court ruling or confirmation that leave to file was granted. Copia Institute — complete account by Cathy Gellis · OpenAI copyright litigation — background

November 1, 2026

• Runway must face creators’ YouTube AI-training circumvention claims — In an October 30 order, U.S. District Judge Lewis A. Kaplan denied Runway AI’s motion to dismiss the consolidated Ace Cam and Businessing complaint. The court held that creators plausibly alleged circumvention of YouTube access controls under the DMCA; public availability of the videos did not by itself defeat those claims. The ruling allows the case to proceed at the pleading stage and does not decide liability, certify a class or resolve whether AI training generally constitutes fair use. Court — October 30 dismissal order · Businessing/Ace Cam — dismissal ruling
• DOJ and states press Google payment-ban appeal, citing AI competition — In an October 29 reply brief, the Justice Department and plaintiff states asked the D.C. Circuit to vacate and remand the district court’s refusal to impose a ban on Google’s search-distribution payments. They argue that the remedy should address default-distribution and revenue-sharing advantages that can also impede rival generative-AI products, and that the district court applied the wrong remedial framework. These are the government plaintiffs’ arguments on cross-appeal, rather than an appellate ruling or an adopted payment ban. DOJ and states — complete cross-appeal reply
• Quad outlines more than $85 million in planned trusted-AI support — In their October 31 joint statement after meeting in Atlanta, the foreign ministers of Australia, India, Japan and the United States expressed their intent to provide more than US$85 million in technical and financial assistance for access to trusted AI across the Indo-Pacific, including through a Quad accelerator. The statement uses “SuperIntelligence,” with a footnote explaining that Australia, India and Japan use the term to refer to artificial intelligence. The announcement states an assistance intention; it does not establish that the money has been disbursed or create a binding AI law. Quad foreign ministers — complete joint statement
• Brown researchers flag legal gaps in emergency AI shutdowns — An October 28 white paper led by Jeff Colgan at Brown University’s Climate Solutions Lab argues that U.S. authorities cannot count on quickly stopping an out-of-control AI system by cutting its electricity. The authors identify uncertain legal authority, court-process delays, difficulty locating the relevant servers, backup power and an unclear chain of command as obstacles when operators do not cooperate voluntarily. Their analysis assumes a dangerous system running on identifiable hardware that cannot immediately relocate. These are the researchers’ conclusions about emergency preparedness, rather than a court ruling or a new shutdown power. Brown CSL — complete October 28 white paper
• Ninth Circuit declines to pause limits on AI-assisted voter-fraud operations — On October 30, the Ninth Circuit denied a government request to stay temporary limits on certain DHS election-related field operations involving people officials know or reasonably suspect are U.S. citizens in nine states. In California Nurses Association v. Mullin, the panel recounted district-court findings that DHS used iLab AI tools to build target profiles from data that could misidentify citizens. The panel found the government had not shown likely irreparable harm from the injunction, which expires when polls close on November 3. Same-day en banc review was declined, with seven judges dissenting on standing, statutory and executive-authority grounds. This is interim stay litigation, not a final ruling on the legality of DHS’s AI tools or all voter-fraud investigations. Ninth Circuit — complete panel order · Court — en banc order and dissent
• Indigenous artists warn AI copies are eroding cultural rights — In a November 1 report, the Guardian quotes Pitjantjatjara artist Sally Scales and Indigenous-art advocates warning that AI scraping, alteration and unattributed imitation undermine artists’ control over cultural knowledge and income. The National Association for the Visual Arts points to these concerns in a submission to Australia’s parliamentary AI inquiry. Researchers Kathy Bowrey and Oliver Bown propose disclosure, creator refusal rights, a levy and takedown mechanisms; these are advocacy proposals, while the attorney general’s office says consultation on copyright options continues. The Guardian — artists and copyright-policy report
• FCA consults on AI-related risks in financial-crime guidance — The UK Financial Conduct Authority’s October 23 consultation proposes updating its Financial Crime Guide to reflect AI-related risks, alongside changes concerning fraud, data security, sanctions and other financial-crime controls. The regulator says firms’ increasing use of AI and third-party tools is among the developments behind the review. Comments are requested by December 24, 2026, before feedback and final guide text are published. This is a consultation on guidance, not newly adopted AI rules. FCA — complete CP26/38 announcement
• Suno asks to pause labels’ second AI-training copyright suit — In an October 30 motion in Massachusetts federal court, Suno asked to stay the record labels’ second lawsuit, No. 1:26-cv-14275-FDS, until summary judgment is resolved in their earlier action. Suno argues that overlapping fair-use and YouTube access-control issues could resolve or narrow claims involving more than 60,000 additional recordings and avoid duplicative discovery. The motion certifies that the plaintiffs oppose the request. It is Suno’s requested pause and legal argument, not a court order granting a stay or deciding fair use. Court filing — complete stay motion · Suno — complete supporting memorandum · UMG/Sony–Suno — stay request
• Travis County rules cover AI courtroom recording — Amended local rules for the Travis County Civil and Family District Courts take effect November 1. Their definition of recording, broadcasting or photographing expressly includes visual, audio and transcription recording by AI programs. Chapter 14 requires prior court authorization to record inside the courtroom and provides for possible contempt or removal for violations. The definition of courtroom includes in-person and virtual proceedings. Travis County — filed local rules, pp. 4 and 22 · Travis County — courtroom recording rules
• National Grid agrees land deal beside planned AI datacentre — National Grid agreed last week to buy 29.5 hectares at Dorman Point on the Teesworks estate for a major electricity substation, the Guardian reports November 1. The infrastructure could supply a neighbouring planned AI datacentre, for which planning applications have been submitted. The deal advances the UK’s AI-infrastructure ambitions following ministers’ publication of a draft national policy statement for datacentres and amid opposition to developments elsewhere. The Guardian — Teesworks land and planning report
• Clients press law firms to share AI efficiency savings — In a November 1 New York Times report carried by the Seattle Times, clients are pressing law firms to explain how AI-related time savings affect their bills. Legora’s Kyle Poe says in-house teams often request reductions of 20% to 30%; firms respond that AI has not necessarily saved that much time and can shift work toward substantive analysis. Ropes & Gray’s Neill Jakobe describes discussing AI use with clients at the start and end of projects and greater openness to fixed fees. These are reported client demands and firms’ responses, rather than measured industry-wide savings or a new billing rule. The New York Times — Sarah Kessler, via Seattle Times
• Alphabet shareholder challenges AI-training oversight in derivative suit — Phil Rosen’s October 30 derivative complaint in Chicago federal court accuses Alphabet officers and directors of breaching their duties over AI training on allegedly unauthorized copyrighted works and voiceprints, Music Business Worldwide reports. The suit seeks compensation for Alphabet and governance reforms; its allegations have not been adjudicated. Music Business Worldwide — complete report · Rosen v Pichai

More months in 2026 · All news years