California's legislature has approved a first-of-its-kind state law setting rules for how lawyers may use generative artificial intelligence in their work. The bill, SB 574, passed both chambers as of Monday and now heads to Governor Gavin Newsom for signature or veto.
The measure prohibits lawyers from delegating the practice of law to generative AI. It requires attorneys to take "reasonable steps" to verify the accuracy of all AI-produced materials, including case citations, and to correct any false or hallucinated output before using it. Lawyers must also disclose AI use in court filings and are barred from entering confidential or nonpublic information into certain generative AI systems.
What the bill requires
Under the legislation, arbitrators cannot delegate decision-making to an AI tool. The bill builds on existing obligations under California's civil litigation rules, such as the requirement that court filings be supported by existing law. It was modeled after a separate rule for judges and court employees adopted by the California Judicial Council.
Senator Tom Umberg, a Democrat who chairs the California Senate's judiciary committee, introduced the bill. He said lawyers continue to "submit materials that have hallucinations or have some other anomaly, and they attribute it to AI, and that simply can't exist."
"Our system, our courts, our judiciary has to rely on the integrity of the litigants and the advocates. And if they can't do that, everything breaks down," Umberg said.
Enforcement and reaction
The verification requirement "should raise the standard for lawyers," Umberg said. The measure also "gives the courts another tool if lawyers don't follow the law, in terms of potential sanctions," he added.
Some California lawyers argue the bill largely mirrors obligations attorneys already face. Wayne Stacy, executive director of the Berkeley Center for Law and Technology at the University of California, Berkeley School of Law, said the measure is mostly "duplicative" of existing ethics rules covering lawyers.
The bill arrives as judges across the country have sanctioned litigants for including erroneous or fabricated AI-generated material in court filings without verifying the output. Professionals working with AI for Legal applications have had to navigate a patchwork of court rules and ethics opinions. For government attorneys and agency counsel, the California measure signals a formalized standard that could influence other states.
Why this matters for government professionals
State and local government attorneys should track this bill closely. If Newsom signs it, California will set the first statutory baseline for AI use in legal practice, creating a compliance benchmark that other jurisdictions may copy. Government legal departments that use generative AI for drafting, research, or case preparation should review their internal policies now against the bill's core requirements: verify AI output, disclose its use, and never feed confidential information into unapproved systems. Policy staff involved in AI governance may also find relevant frameworks in AI for Policy Makers training.
Your membership also unlocks:
