TL;DR
California Governor Gavin Newsom signed an executive order on September 18, 2026 mandating accelerated independent oversight of frontier AI companies and directing experts to design a verified emergency "kill switch," positioning the state as the de facto national AI regulator while Washington stays silent.
What happened
- Newsom's executive order directs the California Government Operations Agency to accelerate implementation of SB 813 and AB 1405, both signed the prior week.
- SB 813 (McNerney) creates a framework for certifying independent verification organizations to audit AI systems; AB 1405 (Bauer-Kahan) establishes a state registry for AI auditors with independence and transparency standards.
- A panel of national experts has two months to recommend further legal changes, including onsite auditor embeds at frontier AI labs.
- Key proposals on the table: mandatory emergency shutoff (kill switch) for frontier models, with ongoing third-party verification of its effectiveness.
- The Hugging Face attack is cited as a triggering incident; the order expands the definition of critical safety incidents to include loss-of-control events.
Why it matters
- No federal AI incident-reporting law exists: the order explicitly fills a vacuum left by the Trump administration and Congress, which have enacted zero AI safety requirements.
- Frontier AI companies operating in California, home to most major AI labs, face the prospect of permanent onsite auditors and a government-verified kill switch, raising compliance costs and operational constraints.
- California's framework (covering frontier safety, deepfakes, children's chatbots, privacy, cybersecurity, and government deployment) is now being pitched as the national regulatory floor, which would extend its reach far beyond state borders.
- The kill switch requirement, if adopted, would be the first government-mandated hard-stop mechanism for frontier models anywhere in the US, setting a precedent other states and eventually Congress could follow.
- Industry faces a two-month countdown before expert recommendations land, compressing the window for lobbying or shaping the final rules.
What to watch next
- Expert panel recommendations in two months: whether the kill switch and onsite auditor proposals survive or get softened will signal how aggressive the final rules are.
- Federal response: any move by Congress or the Trump administration to preempt California's framework, or conversely to adopt it as a national baseline, would dramatically shift the regulatory landscape.
- Frontier lab compliance posture: watch whether major AI companies (headquartered in California) publicly accept, challenge, or relocate operations in response to onsite audit requirements.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.