AI × Law

Norm AI announced it raised $120 million in Series C funding at a $1.2 billion valuation, bringing its total raised to over $260 million. The company's platform automates first-pass legal and compliance tasks and powers Norm Law, an AI-native law firm that uses outcome-based pricing.

At least a dozen U.S. civil and criminal cases have cited subpoenaed or retrieved ChatGPT logs, establishing that AI chat histories carry no privilege or legal confidentiality protections. Courts are treating these logs as standard discoverable electronic records, though judges remain divided on whether AI-assisted legal work qualifies for work-product protections.

Midyear reports from Cornerstone Research and NERA show federal securities class action filings hit a six-year high in H1 2026, fueled by AI-related claims. While AI cases represented just 13% of total filings, they accounted for 73% ($1.3 trillion) of total maximum investor losses, vastly expanding exposure for tech companies and data center operators.

The Seventh Circuit upheld the dismissal of production and possession charges against a defendant accused of creating AI-generated child sexual abuse material that does not depict real individuals. Writing for the panel, Judge John Z. Lee noted that while advanced generative AI models make virtual imagery virtually indistinguishable from real children, the court is bound by Supreme Court precedent in Stanley v. Georgia and Ashcroft v. Free Speech Coalition.

A new legal resource tracks global case law and court guidance regarding the use of AI in expert evidence. The tracker monitors developments including AI-assisted expert reports, disclosure disputes over prompts, and admissibility challenges involving synthetic outputs.

The Arizona Supreme Court has rejected a proposal to ban judges from using artificial intelligence. Instead, the court opted to continue testing AI technology under supervision and newly established rules.

A retired Los Angeles County Superior Court judge has raised concerns over courts deploying 'Learned Hand,' an AI tool generating first drafts of tentative judicial rulings. The commentary highlights ethical and constitutional risks regarding judicial independence, human judgment, and court funding of private AI testing.

The Arizona Supreme Court denied a petition to ban judges from using generative AI for core judicial work through 2029. Instead, the court will continue testing approved AI tools for administrative efficiency and reducing backlogs, while maintaining that judges cannot use AI to draft final legal rulings.

The New York City Bar Association released guidance addressing the ethical use of AI tools for attorneys. The committee warned against recording non-clients without strategic justification, outlined document complexity factors for AI drafting, and cautioned lawyers against inflating bills or charging for saved time when using AI efficiencies.

A comprehensive research profile details the growth of legal AI platform Legora, which has raised $865M in total funding to date and grown its headcount to 700. The company offers an agentic AI system for document review, drafting, and research embedded directly into Microsoft Word and Outlook.

Latin America's largest legaltech platform, Jusbrasil, has acquired MinutaIA, a generative AI drafting and research startup serving over 1,000 law firms and 180 public institutions. MinutaIA will continue operating independently while integrating its workflow tools into Jusbrasil's massive legal data platform.

Harvey has partnered with PacerPro to stream real-time federal docket data directly into its legal AI platform. The integration allows litigators to automatically pull filings, track dockets, and process court data within their AI workflows.

Legal technology startup Norm has raised $120 million in Series C funding led by Khosla Ventures. The round elevates the company to unicorn status with a $1.2 billion valuation.

A growing market of tech tools, including newly launched platforms like CiteSentinel and CiteCheck AI, is targeting AI-generated errors in legal filings. Designed to catch fake cases, miscited statutes, and altered record facts before submission, these tools also give litigators a automated way to check opposing counsel's work.

Croke Fairchild Duarte & Beres LLC has appointed partner David Lopez-Kurtz as its first chief of artificial intelligence and technology innovation. The move reflects the growing trend of mid-sized law firms appointing dedicated leadership to oversee AI adoption and strategy.

Plaintiff Daniel Benjamin Gilbert filed a motion to dismiss Anthropic's fair-use counterclaim alongside a unilateral covenant not to sue over authorized training copies of his book. The novel maneuver attempts to force Anthropic's fair-use defense to focus solely on training data acquired from alleged pirate sources.

The Computer & Communications Industry Association, NetChoice, and A.I. Progress filed a joint amicus brief in 'In re Mosaic LLM Litigation' in the Northern District of California. The groups argue that training generative AI models on copyrighted material constitutes transformative fair use under copyright law.

A Connecticut state court judge sanctioned a pro se plaintiff for embedding invisible, white-colored text in a pleading designed to manipulate AI models into agreeing with his legal arguments. The court revoked the plaintiff's electronic filing privileges, warning that using concealed text injections to influence AI tools disrupts judicial processes.

Justice Jerod Tufte presented to the Big Muddy Bar Association on the North Dakota Supreme Court's approach to AI, emphasizing careful use and scrupulous verification after the Court sanctioned a pro se litigant $500 for submitting fictitious cases in City of Dickinson v. Helgeson. He also detailed Administrative Order 36, which authorizes internal court use of generative AI under specific data-retention and sensitivity guidelines.

India's government has introduced draft guidelines regulating the use of AI in courtrooms. The proposed rules explicitly prohibit algorithmic decision-making, bail or recidivism risk scoring, and party profiling, ensuring final determinations remain solely with human judges.