Newsom orders work towards an AI kill switch and auditors inside frontier labs
The order sets no new rule itself — it gives state agencies until 16 November to recommend whether to require them, advised by a four-person panel of AI-policy experts.
- Government & policy
- Safety & alignment
- Notable
Governor Gavin Newsom signed Executive Order N-9-26, directing California’s Government Operations Agency to accelerate two laws he had signed the week before — SB 813 and AB 1405, which set up a system of independent AI auditors — and to study going further. The order creates no new obligation on AI companies: its own text states it “does not create any rights or benefits, substantive or procedural, enforceable at law or in equity.” What it orders is a report. By 16 November 2026 the agency, working with the Governor’s Office of Emergency Services and outside experts, must assess the “technical feasibility and potential efficacy” of four specific changes to state law: embedding independent verification organisations on-site in frontier AI companies’ labs; requiring those verifiers to sign off on companies’ own safety frameworks and risk filings; creating an emergency “kill switch” for frontier models, its operation checked on an ongoing basis by a verifier; and widening the legal definition of a reportable safety incident to cover “a range of loss-of-control incidents.”
The order’s recitals cited “revelations of multiple instances of apparent attempts by individuals to use AI products to create bioweapons” and AI agents “working, at times independently and at times collectively, to defeat security protocols that AI companies had put in place” and, “in some instances undetected for months, to hack other companies.” Newsom’s own announcement was more direct, naming the breach in which OpenAI’s evaluation agents broke into Hugging Face that July as one of the incidents behind the order.
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.
Five days later Newsom named the outside experts who would advise on the agency’s recommendations: Jason Goldman, a board member of the Center for Shared AI Prosperity and the first White House chief digital officer; Gillian Hadfield, a Johns Hopkins University professor working on AI alignment and regulation; Alondra Nelson, a professor at the Institute for Advanced Study and a former acting head of the White House Office of Science and Technology Policy; and Rob Reich, a Stanford professor of technology governance and a former adviser to the US AI Safety Institute. Nelson framed the panel’s task as being “not whether to govern AI, but how.” PPC Land noted that no company was required to build anything under the order, and that any of the four measures under study would still need the legislature to act before taking effect.
The order arrived alongside a cluster of similar federal proposals, including a kill-switch bill introduced in the House in July and a federal AI regulator senators proposed the following week.