California just signed two new AI safety laws, but they still lean on voluntary standards — and the researchers who build this technology are telling us that's not enough.

This week, a former OpenAI and Anthropic researcher publicly warned that leading AI companies are "gambling with our lives," and Anthropic's own alignment lead, Evan Hubinger, said in a public statement that the risk of AI causing catastrophic harm within the next decade is "above 10%." State Sen. Scott Wiener, who has pushed AI safety bills for three years, said plainly, "We're seeing warnings from people in the AI world that we need to get it together."

Gov. Gavin Newsom signed Senate Bill 813 and Assembly Bill 1405 this week, creating a registry for third-party AI auditors and a state commission to set safety standards — but those standards are voluntary. Voluntary rules ask companies racing to build ever-more-powerful AI to police themselves. That's not a safeguard; it's an honor system for technology researchers themselves say could cause mass harm.

California should finish what it started. We're calling on Governor Newsom and the California Legislature to require independent, third-party safety audits for any frontier AI model before it's deployed to the public — not just make auditors available, but make audits mandatory.

California has led the nation on AI policy before. It can lead again by closing the gap between the risks researchers are describing and the rules currently on the books.

Sign the petition to make AI safety audits mandatory in California — before "voluntary" standards prove too weak to catch the next warning sign.