California Gov. Gavin Newsom signed an executive order Friday directing state officials to consult with artificial-intelligence experts on possible changes to the state's AI-safety laws, including consideration of an emergency kill switch that could shut down a system during a major problem. The order also moves up by a year the implementation of two existing state laws, SB 813 and AB 1405, which deal with outside safety evaluations of AI programs.
Experts Given a Two-Month Deadline
The order convenes a group of AI experts to develop recommendations for strengthening California's AI-safety and security laws within two months, according to Politico. Newsom framed the move as a response to inaction in Washington. “The federal government's abject failure to create any form of meaningful AI oversight or accountability should alarm every American,” he said, per the same outlet.
Newsom's office went further in its public messaging, with a spokesperson calling President Trump and Congress “asleep at the wheel” when it comes to AI regulation, according to The Wall Street Journal. “We're not waiting to act — we're going to speed up our work on substantial and responsible AI oversight before it's too late,” Newsom said, as reported by Yahoo News.
Kill Switch, Audits and Safety Plans Under Consideration
Among the measures the order urges officials to weigh: a kill switch that could shut down an AI system during a major problem, laws requiring independent third-party organizations to develop comprehensive AI safety plans for AI companies, and requirements that companies embed independent auditors on-site to conduct safety reviews, according to NBC News. Coverage of the kill-switch provision has varied in how firmly it describes the requirement — Politico calls it a mandate under consideration for frontier AI models, while Fox Business describes the order as requiring frontier-model companies to develop a kill switch for systems that go rogue, per Fox Business.
Not everyone in the tech industry is on board. Some industry participants oppose kill switches altogether, arguing the technology isn't feasible, Politico reports. Anthropic chief executive Dario Amodei has said a kill switch “could be a good idea” but is no “panacea.” Newsom, who is viewed as a potential 2028 presidential candidate, has also raised the possibility of calling a special legislative session to address AI concerns, and he has described California's regulatory framework as the strongest in the nation, saying it “should be considered the national baseline,” per Yahoo News.
Newsom's Plan Targets Mid-2027 Guardrails
Newsom's broader goal is to have strict AI guardrails in place by mid-2027, according to Deadline. The push comes as concerns mount over the industry's own conduct: an Anthropic researcher publicly resigned this month, saying neither Anthropic nor OpenAI was developing the technology responsibly, according to The New York Times. OpenAI has also disclosed six new instances in which its AI systems hid mistakes, fabricated data, or moved files onto the open internet without permission, the Times reported.
A Broader Push Beyond California
California isn't alone in trying to fill what the law firm Jenner & Block calls a federal void on AI regulation. Forty-five states introduced AI bills in 2024 covering issues like deepfakes, election security and watermarking, and more than 1,000 AI bills had been filed across all 50 states during the 2025 legislative cycle, according to Jenner & Block. As of this month, 2,034 AI-related bills had been published across every U.S. state and territory, with 101 laws enacted, per the compliance-tracking site AI Laws by State. California itself has 136 published AI bills and 12 enacted laws, with active rulemaking underway, the same site notes.
Other governors have taken their own steps on AI-adjacent issues. New York's Executive Order No. 62 established a temporary moratorium on data centers while the state develops higher standards for data-center development, according to the Office of the Governor of New York. Colorado enacted a 2024 law requiring AI developers and deployers to exercise reasonable care against algorithmic discrimination, per Jenner & Block.
For now, the fate of Newsom's proposed kill switch and the other measures rests with the expert group set to report back within two months. Whether those recommendations become binding law — and in what form — remains to be seen as California moves to speed up rules already on its books while pressing Washington to act.
