Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai In Legal Practice topic
No spam. Unsubscribe anytime.
Assembly committee hears SB 574 to require AI transparency in legal filings and decisions
Summary
Senator Umberg’s SB 574 would require attorneys, judges and arbitrators to disclose use of AI in filings, decisions, and arbitration and keep a human ultimately responsible. Proponents said the bill addresses inaccurate AI citations and bias; committee discussion acknowledged scope and enforcement questions.
Senator Umberg presented SB 574 to the Assembly Privacy and Consumer Protection Committee as a transparency measure for the use of artificial intelligence in legal practice, saying the bill would "add transparency to the use of artificial intelligence" when attorneys or judges rely on model outputs.
"Artificial intelligence, of course, is present in all aspects of our life including in the law," the sponsor said, and argued that "it cannot supplant human beings and cannot supplant judgment." He and witnesses urged requirements for attorneys to check citations, remove biased or sensitive personal data, and ensure a human makes final decisions. "If a lawyer sees obvious bias, they have to eliminate" the AI‑generated material, he said.
Supporters in the room — including a witness who identified herself as Tracy Rosenberg of Oakland — said the bill is a necessary first step. The chair noted bar guidance is already circulating to lawyers and framed the measure as harmonizing professional responsibility with new technology. Committee members pressed on definitions and on how courts will detect undisclosed AI use, and discussed whether the bill should specify enforcement mechanisms or rely on existing professional disciplinary processes.
At the hearing close, the committee put the item on the calendar to be moved at the appropriate time pending a quorum and potential motions. The bill drew broadly favorable comments from members who said they wanted to avoid undermining legitimate uses of AI while adding accountability.
