AI
California Just Sent Auditors Inside the AI Labs
On Friday, California Governor Gavin Newsom signed an executive order that puts the state's AI safety ambitions on a faster clock. The order directs the California Government Operations Agency to accelerate implementation of recently enacted AI oversight laws and, together with the Governor's Office of Emergency Services, convene national experts to recommend additional safeguards for advanced AI systems. The expert group has two months to deliver its guide.
The proposals on the table are concrete. Companies developing advanced "frontier" AI models could be required to build an emergency shutoff mechanism, the so called kill switch, with its effectiveness subject to ongoing verification by an independent organization. Frontier AI companies could be required to host independent verification organizations onsite, inside their labs, conducting regular audits and evaluations. The state is also considering independent verification of the safety frameworks, transparency reports and safety assessments that AI companies must file, plus an expanded definition of critical safety incidents that includes loss of control events.
The order builds on laws Newsom signed last week. Senate Bill 813 establishes a framework for certifying independent verification organizations to assess AI systems for compliance with state law. Assembly Bill 1405 creates a state registry for AI auditors and sets standards for their independence, transparency and integrity. California already requires frontier model developers to publicly disclose safety frameworks and report specified critical safety incidents under 2025 legislation, alongside laws covering AI auditors, companion chatbots used by children, AI generated content, privacy and state government use of AI.
Newsom framed the move as a direct answer to Washington's inaction. In a statement he said California will speed up its work on substantial and responsible AI oversight, calling the state the holder of the strongest AI regulatory framework in the nation and saying its policy should be the national baseline. The backdrop supports his case: the House adjourned for midterm campaign season leaving AI safeguards on the table, Speaker Mike Johnson cautioned against moving quickly, and 107 Democrats led by Representative Sam Liccardo demanded the chamber stay in session to address AI safety concerns.
Whatever one thinks of the politics, the mechanism is new. Independent auditors embedded inside frontier labs, verifying safety frameworks and testing kill switches on an ongoing basis, would give regulators a kind of continuous visibility into the systems they oversee that is entirely new. Two months from now, when the expert guide lands, the country will see whether that model is workable or simply ambitious.
If you build with AI, sell AI, or simply live downstream of the labs, California's rulebook is becoming your rulebook. The state where the frontier models are built is writing the compliance reality, and the companies that design for independent verification now will have a head start when, and the direction of travel suggests when rather than if, those standards travel further. The era of self graded AI safety is closing; the graded era is being drafted in Sacramento.
Quick answers
What is this story about?
On Friday, California Governor Gavin Newsom signed an executive order that puts the state's AI safety ambitions on a faster clock. The order directs the California Government Operations Agency to accelerate implementation of recently enacted AI oversight laws and, together with the Governor's Office of Emergency Services, convene national experts to recommend additional safeguards for advanced AI systems. The expert group has two months to deliver its guide.
Why does this story matter?
If you build with AI, sell AI, or simply live downstream of the labs, California's rulebook is becoming your rulebook. The state where the frontier models are built is writing the compliance reality, and the companies that design for independent verification now will have a head start when, and the direction of travel suggests when rather than if, those standards travel further. The era of self graded AI safety is closing; the graded era is being drafted in Sacramento.
Sources
- MyNewsLA: Newsom orders expanded AI oversight, advances kill switch
- Bloomberg Law: Newsom signs order requiring AI labs to develop kill switch
New to crypto? Read the crypto glossary, browse frequent questions, read our story, or explore the story archive.