Newsom signs AI transparency bill prioritizing safety

SACRAMENTO — Gov. Gavin Newsom signed a bill Monday that will create new transparency measures for large AI companies, including public disclosure of security protocols and reports of critical safety incidents.

Sen. Scott Wiener (D-San Francisco) said Senate Bill 53 will create “commonsense guardrails” to ensure groundbreaking innovations don’t sacrifice safety and transparency amid the rapid growth of AI technologies. Newsom said the bill strikes the right balance of working with the artificial intellegence companies, while not “submitting to industry.”

“AI is the new frontier in innovation, and California is not only here for it – but stands strong as a national leader by enacting the first-in-the-national frontier AI safety legislation that builds public trust as this emerging technology rapidly evolves,” Newsom said in a statement.

The bill was introduced this year after Newsom vetoed a broader bill last year, which was also authored by Wiener. That bill, SB 1047, was supported by Elon Musk and prominent AI researchers, but opposed by Meta and OpenAI.

In his lengthy veto message last year, Newsom called SB 1047 “well-intentioned” but added that it was not the “best approach to protecting the public from real threats posed by the technology.” In punting the measure last year, Newsom announced that his administration would convene a working group of AI leaders and experts to develop more workable protections that became the basis for SB 53.

The new law will require companies to disclose their safety and security protocols and risk evaluations. It mandates reporting of critical incidents — such as cyberattacks or unsafe behavior by autonomous AI systems — to the state’s Office of Emergency Services.

Cal OES would begin publishing annual reports in 2027 that anonymize and aggregate critical safety incidents it receives. SB 53 also strengthens whistleblower protections for employees who report violations.

The Attorney General in California will be able to bring civil penalties of up to $1 million against companies who violate the new law.

“With a technology as transformative as AI, we have a responsibility to support that innovation while putting in place commonsense guardrails to understand and reduce risk,” Wiener said in a statement.

The bill was opposed by the California Chamber of Commerce and the Chamber of Progress, a tech industry association.

“This exhaustive approach compels developers to allocate significant time and resources toward preparing for hypothetical risks rather than addressing actual, demonstrable harms,” wrote the Chamber of Progress.