California Passes AI Rules for the Legal Field
California finds itself leading the way with artificial intelligence (“AI”) regulation after successfully passing a law that would govern legal practitioners’ use of AI. On August 31, 2026, California’s state legislature unanimously passed Senate Bill 574 (the “Bill”), which will serve to impose legal requirements surrounding the use of AI by attorneys, arbitrators, judicial officers, and alternative resolution providers. SB 574 codifies guidance from the California Judicial Council on AI use in legal proceedings.[1] The Bill is now with Governor Newsom for signature.
Generally, the Bill prohibits attorneys from “delegat[ing] the practice of law to generative artificial intelligence.[2] Moreover, any attorney who uses generative AI shall be subject to additional restrictions. First, attorneys using AI must not enter confidential information, personal identifying information, or other non-public information into a generative AI system unless access to that data within the system is restricted to the attorney and to people the attorney has authorized who are themselves bound to protect its confidentiality. Second, the attorney must take reasonable steps to verify that the AI’s output is accurate, including checking the cases and statutory citations it generates, and must correct any errors or hallucinated content before using it in any material. Third, the attorney must disclose the use of generative AI to the court in any document submitted to the court and must also consider whether to disclose its use when it helps create content shared with the public.[3] The Bill also requires that “a brief, pleading, motion, or any other paper filed in any court shall not contain any citations that an attorney responsible for submitting the pleading has not personally verified, including any citation provided by generative artificial intelligence.”[4]
The Bill is not limited to attorneys, however, as arbitrators and judicial officers would separately become subject to additional restrictions and considerations. Arbitrators are prohibited from delegating any part of their decision-making process to any generative AI tool and are further restricted from relying on the output of generative AI outside the record without making appropriate disclosures to the parties beforehand, and, as practical, allowing the parties to comment on such use.[5] Judicial officers currently look to Standard 10.80 of the California Standards for Judicial Administration for applicable guidance for using generative AI, and the Bill would require the Judicial Council to publicly revisit and revise, as necessary, the aforementioned standard to incorporate changes to reflect development in AI.[6]
It is likely too early to determine the impact these laws, and their interpretation, will have on the use of AI in the legal field, but it serves as a good reminder to utilize AI in a responsible and thoughtful manner. We’ll continue to monitor for any significant updates.
[1] https://sd34.senate.ca.gov/news/bloomberg-law-california-attorney-ai-guardrail-bill-advances-newsoms-desk; https://courts.ca.gov/advisory-body/artificial-intelligence-task-force.
[2] S.B. 574, 2025–2026 Reg. Sess. § 1 (Cal. 2026) (proposing CA BUS & PROF § 6068(a)(2)).
[3] Id. (proposing CA BUS & PROF § 6068(a)(3)).
[4] Id. (proposing Cal. Code Civ. Proc. § 128.7(b)(2)).
[5] Id. (proposing Cal. Code Civ. Proc. § 12821.1(b)).
[6] Id. (proposing Cal. Code Civ. Proc. § 180); https://courts.ca.gov/cms/rules/index/standards/Standard10_80.
