AI × Law

This roundup says courts and regulators are drawing sharper lines on generative AI training and copyright, including U.S. rulings for Anthropic and Meta, France’s €250 million fine against Google, and the U.K.’s first AI copyright judgment. It also highlights Anthropic’s $1.5 billion settlement over pirated books, a huge signal for training-data risk.

HAQQ is offering startups a first month free on its legal AI platform, with plans for founders, startups, and portfolio companies through VC and incubator partnerships. The product says it covers startup legal workflows from incorporation and SAFEs to fundraising, contracts, e-signatures, trademarks, and compliance across multiple jurisdictions.

Supio published a 2026 survey of 207 U.S.-based personal injury attorneys and firm leaders finding that 78% have engaged with AI, but only 30% use it regularly. The report says the main blocker is trust and verification, not cost, and highlights growing concern over untraceable AI output and sanctions tied to hallucinations.

Two candidates for the Washington Supreme Court are debating whether judges need new AI-specific rules. One wants a binding prohibition on AI making decisions; the other says existing ethics canons are enough and notes AI is already being used as a research and efficiency tool in courts.

OpenAI has hired former Ironclad CEO Jason Boehmig and is collaborating with Willkie Farr & Gallagher to roll out ChatGPT Enterprise to every attorney and integrate its models into the firm’s tools. The article says OpenAI is now courting lawyers directly, a shift that could put it in more direct competition with legal AI startups and existing vendors.

Law.com reports that firms including Cooley, Fenwick, Sheppard and Pillsbury are using associates to evaluate AI tools for legal work and giving them credits to do so. The article says these firms are tapping associates to help determine how AI fits into practice, reflecting client pressure and a broader push to measure real-world impact.

Wisner Baum partner W. Crawford Appleby said the firm set up an AI committee to review new tools and train lawyers on how to use them. The piece is a practical look at firm-level AI governance and the risk of costly mistakes, but it does not report a major new product, deal, or policy change.

A district court ruled that using AI prompts can waive privilege protection for both litigants and lawyers. The decision is a notable warning for legal teams using AI in their workflows, especially when handling confidential or privileged material.

Scale AI agreed to pay $12.5 million to resolve a California class action alleging worker misclassification, unpaid wages, unreimbursed business expenses, and other labor-law violations. The settlement covers contributors, including generative AI annotators and prompt engineers, and shows the employment risks around AI labor platforms.

The Alabama State Bar updated its ethics guidance to say lawyers using artificial intelligence cannot charge clients for hours saved by the technology, though they can bill for time spent reviewing AI output. It’s a concrete billing-and-ethics signal for firms experimenting with AI tools.

Oliver Roberts of WashU Law presented to the Pennsylvania Conference of State Trial Judges on AI in the courts, covering judicial use of AI, sanctions and court decisions involving improper AI use, hallucinated authorities, evidentiary issues, and judicial ethics. The session also included live demos from Thomson Reuters, LexisNexis, xAI, and Learned Hand, giving judges a look at AI tools for research and court workflows.

A weekly Legaltech Rundown highlights several concrete product moves across the market, including Epiq’s expanded agentic AI in Epiq AI Accelerate, NetDocuments’ new apps for trademark prosecution and plaintiff-side litigation, and Spellbook’s launch in Google Docs. The roundup also notes new partnerships, leadership changes, and funding, showing continued vendor momentum in legal AI.

Clio published a guide on how new law firms can use its AI tools for intake, billing, drafting, research, and case analysis. The piece also says 79% of legal professionals use AI in some capacity, underscoring how quickly AI has become part of firm operations.

RELX said LexisNexis revenue rose 10% to a record £959 million in the first half of 2026, with growth in law firm and corporate legal business driven in part by adoption of its AI products. The company highlighted Lexis+ with Protégé, Protégé Work, and a multi-model strategy, underscoring the scale and staying power of an incumbent legal AI provider versus venture-backed startups.

Connecticut courts have adopted new rules requiring lawyers and self-represented parties to verify any generative AI output used in court filings. The move is notable because the state appellate rules now allow sanctions for failures to check AI-generated citations, and federal judges in Connecticut have also warned that hallucinated filings may trigger Rule 11 sanctions.

Clark Hill said it has expanded its AI platform through an enterprise agreement with Thomson Reuters that gives every attorney access to Next Gen CoCounsel, Westlaw Advantage, and Practical Law. The firm says the move follows more than two years of planning, pilots, governance work, and training, signaling continued Big Law investment in vetted generative AI.

GC AI published a launch-week piece on how lawyers can use ChatGPT and when they should not. The most newsworthy detail is the company’s claim that, as of July 2026, its enterprise legal AI platform is used by 1,900+ legal teams, including 80+ public companies and 25 unicorns such as Snyk, Liquid Death, Columbia Sportswear, and Hitachi.

After 17 years at Quinn Emanuel, commercial litigator Chris Kercher has launched Kercher Law, a boutique built around using AI to handle much of the “context assembly” in litigation. He says the firm will lean into fixed-fee, phase-based billing and use tools like Claude and Syllo AI to reshape how complex cases are staffed and priced.

xAI filed a federal lawsuit in Texas against Terry Wayne Harwood, alleging he used Grok to create sexually explicit deepfakes of real adults and children and repeatedly tried to bypass safety systems. The case is notable because AI companies rarely sue individual users, and it could test how far developers can go in using litigation to police harmful misuse.

A lawsuit filed in San Francisco County Superior Court accuses OpenAI and Sam Altman of using ChatGPT-4o to give a former pastor dangerous medical advice as he developed bilateral pulmonary embolisms. The complaint claims negligence and California consumer protection and privacy violations, making it a notable AI liability case for litigators watching product-safety and AI-harm claims.