California Takes Significant Step Forward To Regulate The Use Of AI By Attorneys And Arbitrators

On January 30, 2026, the California Senate passed a new bill, SB 574, to impose restrictions on attorneys (and arbitrators) using generative artificial intelligence (“generative AI”) in their practice. SB 574 is not yet the law – it will now proceed to Assembly committees for hearings and a potential vote before lawmakers adjourn in late August 2026.

The proposed bill, which was introduced by Senate Judiciary Chair Tom Uberg, D-Santa Ana, reflects a growing concern that the rapid adoption of generative AI creates unique risks in certain regulated professions, such as the law. Even robust, well-trained AI models invent, or “hallucinate,” supporting legal authorities in their drafts. To date, there have been hundreds of cases across the country in which attorneys were caught filing briefs and other documents in court that contain cases, quotes, and citations that were “hallucinated” by AI and do not actually exist.

SB 574 seeks to regulate attorneys' and arbitrators' use of “generative artificial intelligence,” which is defined to mean “an artificial intelligence system that can generate derived synthetic content, including text, images, video, and audio that emulates the structure and characteristics of the system's training data.” The proposed bill would codify existing guidance from the California State Bar on the ethical considerations associated with generative AI use, and impose some additional restrictions on both attorneys and arbitrators.

Here's what legal practitioners need to know about this proposed bill:

• Verify AI Citations. Attorneys would be required to “verify the accuracy” of any material created by generative AI; correct “any erroneous or hallucinated output” in any such material; and remove “any biased, offensive, or harmful content” in any such material. The bill would also add a new requirement that any documents filed with a court shall not contain any citations that the attorney responsible for that document “has not personally read and verified.” In other words, AI may assist with preparing drafts, but attorneys remain responsible for the work product they file.
• Protect Confidential Client Information. SB 574 would prohibit attorneys from entering any “confidential, personal identifying, or other nonpublic information” into “a public generative artificial intelligence system.” This would prohibit attorneys from copying confidential information learned from clients or other litigation participants into public AI platforms. The bill does not define “confidential” or “nonpublic” information, but it does define “personal identifying information” to include birthdates, Social Security numbers, driver's license numbers, financial account numbers, addresses and phone numbers, along with “any other content sealed by court order or deemed confidential by court rule or statute.”
• Prevent Bias. The act would also impose a requirement that the use of generative AI “does not unlawfully discriminate against or disparately impact individuals or communities based on age, ancestry, color, ethnicity, gender, gender expression, gender identity, genetic information, marital status, medical condition, military or veteran status, national origin, physical or mental disability, political affiliation, race, religion, sex, sexual orientation, socioeconomic status, and any other classification protected by federal or state law.”
• “Consider” Disclosing Use of Generative AI. SB 574 would add a new duty for an attorney to “consider” whether to disclose the use of generative AI “if it is used to create content provided to the public.”

SB 574 does not authorize any independent penalties for any violations. Instead, violations of the requirement that an attorney must personally read and verify all citations in documents filed with the court are subject to sanctions determined by that court. Other violations could be addressed through the California State Bar disciplinary process. 

Restrictions on Arbitrators. Finally, the bill would prohibit arbitrators from delegating any decision making to a generative AI tool, and require arbitrators to “assume responsibility for all aspects of an award.” It would also require arbitrators to avoid relying on any information created by generative AI that is outside the record, without “making appropriate disclosures” to the parties and allowing the parties to comment on its use. The bill does not contain any language discussing potential penalties on arbitrators for violating any of these requirements.

One thing is for sure: as attorneys continue to adopt generative AI in their practices, other states are likely to consider regulating that use in some form or fashion.

The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.