California Gov. Gavin Newsom issued an executive order on Friday that puts the state in position to take the lead on AI oversight, including the possibility of requiring a 'kill switch' for frontier models.
The order directs the state to convene a group of experts, who must deliver recommendations within two months on how to strengthen AI safety measures in state law. Among the ideas Newsom wants them to examine: independent verification groups onsite for regular audits, transparency reports and risk assessments subject to independent auditor standards, a 'kill switch' that is routinely verified as effective, and reporting 'loss-of-control incidents' as critical safety events.
The order also tells a state agency to speed up two AI laws Newsom already signed. Those laws create a framework for independent verifiers to assess AI safety and a state registry of AI auditors. In CNBC's account, the enhanced laws may include a requirement for independent third parties to write safety plans for frontier AI companies, while The Verge reported that Newsom also floated the possibility of a special legislative session on AI.
Newsom is presenting the move as a model for Washington. He called on Congress and President Donald Trump 'to review and adopt the state's framework, or use its nation-leading regulation as a floor, not a ceiling, for the benefit and safety of all Americans,' according to The Verge. CNBC quoted Newsom saying California has already built 'a national model' and that 'our policy should be the national baseline.'
The timing is obvious. There has been a flurry of AI bill talk in Congress after a former Anthropic researcher warned that AI could 'kill all humans' within a decade. But both The Verge and CNBC say major federal action looks unlikely in the near term, and Newsom says the federal government has failed to create meaningful oversight or accountability.
That leaves California doing what Washington will not: setting the rules first, then asking everyone else to copy them. For AI companies, that means more audits, more reporting, and a lot less room to sell 'trust us' vaporware as governance.
For everyone else, the principle is simple. If frontier models are powerful enough to justify a kill switch, they are powerful enough to justify real accountability. The question is whether California is building a sane safeguard — or just a new layer of bureaucracy that becomes the de facto national standard by default.
