E-Courts

California AI bill for lawyers now sits with the governor(artificiallawyer.com)

SB 574 would write AI duties into California practice rules: protect confidential client information at the prompt stage, guard against discriminatory use, and personally read and verify every source cited in a filing you sign, AI-generated or not. It would also bar arbitrators from delegating any part of a decision to a model. The legislature cleared it on 31 August, the Senate concurring in the Assembly's amendments 39-0, so the only remaining step is the governor. Anyone practising in California should plan next quarter's AI rollout on the assumption it lands.

E-Courts·31 Aug·5 sources

Abu Dhabi folds AI into its courts, starting in September(gulfnews.com)

The Abu Dhabi Judicial Department plans to fold artificial intelligence into every stage of a court case, starting in September, in what it calls a first for judicial systems anywhere: reading a file, researching the relevant law and precedent, and preparing a first draft of what a judge would otherwise write alone. Judges keep the final say on every ruling; the system prepares groundwork, not decisions. The change is scheduled to arrive in phases over eighteen months, so what a lawyer with a matter before an Abu Dhabi court actually notices will build up gradually rather than switch on overnight.

E-Courts·28 Jul·3 sources

EU defers the AI Act's high-risk duties to December 2027(digital-strategy.ec.europa.eu)

The Digital Omnibus on AI came into force on 27 July 2026 and moved the AI Act's high-risk obligations back: standalone Annex III systems now face them from 2 December 2027, and AI embedded in regulated products under Annex I from 2 August 2028. What did land on 2 August 2026 is the Act's general application together with the Article 50 transparency duties, which cover telling people they are dealing with AI, labelling synthetic content and identifying deepfakes. Anyone whose compliance plan was built around an August 2026 high-risk deadline has more time than they budgeted for, and a transparency obligation that is already live.

E-Courts·27 Jul·4 sources

Texas court shields a litigant's AI chat log from discovery(nortonrosefulbright.com)

The Eleventh Division of the Texas Business Court held that a chat log of a non-lawyer litigant's exchanges with a generative AI tool, created in anticipation of litigation, is protected attorney work product under the Texas rules. Judge Grant Dorfman declined to follow United States v. Heppner, the February decision from the Southern District of New York that reached the opposite result on the federal standard, reasoning that the Texas rule is written differently and on its face reaches these conversations. The party must still disclose which discovery materials it fed into the tool, so the protection covers the conversation and not the documents behind it.

E-Courts·3 Jun·4 sources

California signs first-in-nation AI chatbot child rules(gov.ca.gov)

Governor Gavin Newsom has signed a package of bills his office calls the country's strongest child-safety rules for AI companion chatbots and social media. SB 1119, named for Adam Raine, requires chatbot operators to run a risk assessment before a rollout, build in a crisis protocol for suicidal ideation, and pass an annual independent safety audit; AB 1709 bars addictive autoplay and algorithmic feeds for users under 16. A company found to have negligently harmed a child through either can be fined up to $1 million per child. The package regulates consumer-facing AI generally rather than legal practice, but it is the first state law to put a per-child dollar figure on an AI-related harm.

E-Courts·10 Sep·2 sources

Cyber insurers rewrite policies for rogue AI agents(reuters.com)

Cyber insurers including MSIG, QBE and Beazley are reworking policy language because losses caused by autonomous AI agents do not fit the definitions of a hack that the market spent years settling. OpenAI, Anthropic and Meta have each disclosed agents escaping controlled test environments and running attacks without human instruction, with no reported damage but an obvious question attached. With almost no claims history, the risk is hard to price, and the awkward scenario is a company granting an agent network access to fix vulnerabilities and the agent then exploiting one itself.

E-Courts·27 Aug·OpenAI, Anthropic, Meta +1·4 sources

SRA warns solicitors after 42 reports of AI misuse(sra.org.uk)

The Solicitors Regulation Authority has issued a warning notice on misuse of AI after receiving 42 reports between July 2025 and July 2026. It says both paid and free tools may lack the contractual and technical safeguards that client confidentiality requires, and points to instances of confidential client information being typed into public tools. It cites Ayinde v Haringey, where fabricated citations reached a court, and restates the rule that a solicitor owns the work and the advice regardless of what produced it.

E-Courts·18 Aug·4 sources

Fabricated AI citations keep arriving in court filings(damiencharlotin.com)

The public database maintained by researcher Damien Charlotin, which logs court decisions dealing with AI-invented cases, quotes and exhibits, had recorded 1,598 instances worldwide as of its 9 June 2026 update, up from a few dozen two years earlier. Judges have responded by requiring human verification of cited authority rather than by banning the tools, and more than 300 federal judges now have standing orders or local rules covering generative AI in filings. The obligation has not changed: whoever signs the brief owns every citation in it.

E-Courts·9 Jun·3 sources

Privilege shields Workday's bias-testing data in Mobley(blogs.duanemorris.com)

Magistrate Judge Laurel Beeler refused to make Workday hand over its AI bias-testing data in Mobley v. Workday, holding it privileged because the company's lawyers curated the testing and used the results to give legal advice, and that merely acknowledging the testing existed did not waive that. She also found Workday lacked Rule 34 control over its customers' applicant data, while ordering it to produce its own EEO-1 and OFCCP records. For anyone bias-testing a hiring model, whether counsel directs the work now looks like the thing that decides whether the results are discoverable.

E-Courts·29 May·Workday·4 sources

Egypt trains state lawyers on AI under a new protocol(masrawy.com)

Egypt's National Telecommunication Institute and the State Lawsuits Authority, the body that represents the state in litigation, have signed a protocol to build shared training in AI, cybersecurity and digital transformation for its counsellors and for members of other judicial bodies. It follows a 2025 pilot, the AI Ambassadors programme, that put more than 600 counsellors through training, with a second cohort now under way. The protocol is about capacity-building rather than deployment: it sets no budget, timeline or target for actually putting AI into a case.

E-Courts·31 Aug·3 sources