Key Takeaways
- California’s SB 574 imposes broad duties on attorneys to disclose the use of generative AI in court filings. However, the law leaves much undefined, so its practical effect remains to be seen.
- Attorneys practicing in California should be aware of the verification and disclosure requirements that take effect on January 1, and those practicing in New York will want to monitor the progress of Bill S2698.
On September 30, California Governor Gavin Newsom signed into law a bill that legislates standards for how attorneys may use generative AI in legal work, effective January 1, 2027. SB 574 states that lawyers “shall not delegate the practice of law” to generative artificial intelligence, and that lawyers must take “reasonable steps” to verify the accuracy of AI-generated materials, such as case citations, and to correct “erroneous or hallucinated” material. The statute does not define what constitutes delegating the practice of law nor what constitutes “reasonable steps” to verify citations. Nevertheless, violations of SB 574’s citation-verification requirement can result in sanctions against counsel and their firms, with represented clients shielded from monetary sanctions.
SB 574 also prohibits inputting confidential or private information into AI tools where access to such inputs is not restricted to an attorney or those authorized by the attorney and prohibits arbitrators from delegating legal decision-making to AI systems. The bill requires attorneys to disclose the use of generative AI in court filings but does not prescribe the form of disclosure nor any required details. The bill is the first to set a statewide statutory requirement for disclosure of generative AI use in court filings, although a similar legislative proposal pending in New York would require a separate certification of filings produced using generative AI.
The proposed New York bill, Senate Bill S2698, if enacted, would specifically require the brief of an appellant to contain a disclosure of the use of generative AI in the drafting of the brief and certification that the content therein was reviewed and verified by a human. Like SB 574, S2698 does not designate a disclosure’s required form or details.
As we have previously reported, several states have acted to regulate the use of AI in licensed professions. California’s new legislation and New York’s proposed certification requirements similarly address attorney responsibilities regarding use of generative AI and may prompt other states to clarify or tighten limits on AI use in legal contexts. California’s SB 574 and New York’s proposed Senate Bill S2698 suggest that verification and disclosure obligations may increasingly define responsible generative AI use in the legal profession.
