California sets new rules for lawyers and generative AI

SB 574 bars attorneys from delegating the practice of law to generative AI.[3] Attorneys must verify AI outputs, correct erroneous or hallucinated material they use, and disclose AI use to courts for all documents submitted to them.[3] For public-facing content generated with AI, attorneys must consider whether disclosure is appropriate.[3] The law defines generative AI as a system that generates synthetic content emulating the structure and characteristics of its training data.[3] These obligations make clear that using a tool does not transfer responsibility for the work to the tool.[3]

An attorney responsible for a court filing must personally verify every citation in it, including citations provided by AI.[3] The statute also bars attorneys from entering confidential, personally identifying or other nonpublic information into an AI system unless access is restricted to authorized people bound by confidentiality obligations.[3] Its examples of personally identifying information include birth dates, Social Security numbers, medical information and financial information.[3] Teams can use these requirements to review filing and data-handling procedures before the law takes effect, with particular attention to citation checks and the information entered into AI tools.[1][3]

The law bars arbitrators from delegating any part of their decision-making process to generative AI.[3] Arbitrators also may not rely on AI-generated information from outside the record without first making appropriate disclosures to the parties.[3] The law directs the Judicial Council to revisit a judicial-administration standard to account for further development of generative AI.[3] It also requires procedures for the State Bar or an approved professional organization to receive, investigate and resolve complaints about ethical compliance by certified firms, providers or practitioners.[3] The alternative-dispute-resolution certification program does not require applicants to be State Bar licensees.[3]

Reuters described SB 574 as the first state law specifically regulating lawyers’ and arbitrators’ use of generative AI, and reported that it builds on existing state rules governing lawyers’ work in civil litigation.[2] A&O Shearman partner Daren Orzechowski said the law overlaps with existing ethics rules but reminds lawyers to check and verify AI output.[2] Baker Botts partner Ariel House said personally verifying citations could substantially increase review time, especially for solo practitioners and small firms.[2] Farella Braun + Martel trial lawyer Jeff Fisher said he did not view the law as a major change to lawyers’ expanding AI use, adding that AI’s benefits and efficiencies are too significant to ignore.[2] Reuters also reported that courts around the country have sanctioned or warned lawyers for failing to vet AI-generated filing material.[2]