California Proposed Law on Gen AI For Lawyers

Proposed California Law Bridges the Gap Between AI Innovation and Attorney Ethics

It was only a matter of time before state legislatures addressed lawyers' use of “Generative AI” in litigation. With its ever-growing use in law, there have been dozens of orders calling out AI-hallucinated case citations from the use of Gen AI. (An often cited case for this proposition is Sanders v. United States, 176 Fed. Cl. 163, 167 (2025)). 

In December 2025, the ugly use of Gen AI appeared with fabricated facts in a declaration that manufactured citations to deposition transcripts, grossly mischaracterized other testimony, and misstated facts in the record, in an opposition to a motion for summary judgement. Pauliah v. Univ. of Miss. Med. Ctr., No. 3:23-CV-3113-CWR-ASH, 2025 U.S. Dist. LEXIS 267306, at *7-8 (S.D. Miss. Dec. 30, 2025). 

Gen Ai making up facts in a motion for summary judgment can result in finding a declaration was filed in bad faith, loss of the motion, and sanctions. 

California’s Response to Improper Use of Gen AI in Litigation 

Lawyers have a duty of candor to courts and a duty of competency for practicing law. Surrendering those ethical obligations begins with sanctions orders from courts. State legislatures are now taking action to prevent lawyers from improperly using Gen AI in court. 

California is considering passing a new bill, SB 574, to address the use of Gen AI by lawyers and arbitrators to add Generative AI to Business and Professions Code Section 6068.1. 

The proposed law defines “Generative artificial intelligence” as “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.” 6068.1(b)(1). 

The proposed additions require lawyers to take “reasonable steps” over the use of Gen AI:

• Verify the accuracy of all generative artificial intelligence (AI) material, including content prepared on their behalf by others.
• Correct any erroneous or hallucinated output in any generative artificial intelligence material used.
• Remove any biased, offensive, or harmful content from generative artificial intelligence material, including content prepared on their behalf by others.

6068.1(a)(3)

The legislation would also amend the California Code of Civil Procedure Section 128.7 to include section (b)(2), which states: 

• A brief, pleading, motion, or any other paper filed in any court shall not contain any citations that the attorney responsible for submitting the pleading has not personally read and verified, including any citation provided by generative artificial intelligence.
• For purposes of this paragraph, “generative artificial intelligence” means 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 legislation also prohibits arbitrators from using Gen AI as part of their decision-making process in proposed California Code of Civil Procedure Section 1282.1(a)(1). 

Why This Matters to eDiscovery 

The proposed amendments get to the heart of a lawyer’s ethical duties for filing accurate citations with the court. We are used to fake case authority headlines. However, there is now a documented case of hallucinated factual support in a declaration illustrating other ways “Gen AI” raises ethical concerns for attorneys. 

Gen AI can be an amazing tool for lawyers. It can expedite first-pass document review. It can help lawyers identify records supporting arguments in motion practice, whether it is undisputed facts in a motion for summary judgment, or to exclude evidence in motions in limine, or any of the many other motions attorneys can bring before the court. 

Gen AI tools in eDiscovery applications (so far) have helped with enhanced search abilities and summaries of records. Motion practice with citations from Gen AI summaries of ESI is a real concern as the technology continues to be adopted, because any factual assertion to a court has to be correct. 

Lawyers need to ensure they maintain their duty of candor to the court. This arguably already includes that the lawyer has actually verified any citation in a brief, pleading, or motion. To ensure this practice, California is considering making it abundantly clear that lawyers must verify the accuracy of any information identified by Gen AI for any court filing. 

That would include evidentiary support that Gen AI can identify. 

The lesson for attorneys is clear: read before you cite.