California's SB 574 Codifies First Statutory Rules on Lawyers' GenAI Use
Signed Sept. 30 and effective Jan. 1, 2027, SB 574 is the first U.S. statute governing attorneys' generative AI use, requiring verification, confidentiality compliance, disclosure and personal citation checking in California filings.
California's SB 574, signed into law Sept. 30 and effective Jan. 1, 2027, is the first state statute expressly governing lawyers' generative AI use. It reaches California-licensed attorneys, counsel appearing before California courts, and arbitrators subject to California arbitration procedures. Duties include not delegating the practice of law to AI, reasonably verifying AI output, correcting hallucinated content, honoring new confidentiality limits on entering nonpublic or personally identifying information, and satisfying disclosure obligations. The provision counsel should flag internally: attorneys must personally verify every citation in filings submitted to California courts, regardless of whether the authority came from AI, a colleague or a research platform. The statute's broad definition leaves open whether AI features inside legal research and document review tools are covered.
For major-firm clients this is a governance and documentation mandate. The alert pairs the statute with 2026 decisions in U.S. v. Heppner (S.D.N.Y.), declining privilege protection for communications with a public AI tool, Warner v. Gilbarco (E.D. Mich.), preserving work product on different facts, and an expert-discovery ruling in Conservation Law Foundation v. Shell (D. Conn.) — suggesting prompts and outputs can surface in discovery. Next step: audit AI policies, vendor confidentiality terms, discovery protocols and expert engagement letters before January 2027.