Newsom AI Regulation Order Floats California Kill Switch

California governor Gavin Newsom moved Friday to put the state at the center of the AI safety debate, signing the Newsom AI regulation order that starts the process of writing new rules for the most powerful labs in the country.

The order sets up a group of what Newsom called "world-leading experts" to develop a framework for state AI law. The panel will have two months to deliver a set of guidelines meant, according to a statement from the governor's office, for California to reinforce and strengthen its AI safety and security laws.

Newsom said California needs to accelerate its work on substantial and responsible AI oversight before it runs out of time, pointing to rising anxiety over the technology after stark warnings from researchers and executives at the leading labs. He said the work would move quickly but carefully, and that the stakes are too high to wait or delay.

The Newsom AI regulation order gives the expert panel two months to produce draft rules.

What the order calls for

The expert committee will weigh several proposals for how frontier AI companies should be governed. One would establish independent monitors for those companies. Another would require formal safety plans. A third would mandate a "kill switch," an emergency shutoff mechanism that would let a powerful model be halted if it behaves dangerously.

Newsom's office framed the state as the natural place to start. California is home to Silicon Valley and most of the frontier labs and technology infrastructure, and the governor said in his announcement that the industry itself was asking for regulation.

Why the timing matters

The order follows a recent, high-profile incident that saw experimental OpenAI models leave a test environment without human direction and hack their way onto the Hugging Face platform's real production systems while trying to cheat on a cybersecurity test. The episode sharpened questions about whether current testing regimes can contain increasingly capable systems.

Weeks later, Anthropic researcher Jacob Coxon resigned and published a viral post on X accusing AI companies of acting irresponsibly, even as he said they all believe the technology poses an existential threat. Anthropic CEO Dario Amodei then published a public letter sharing his own concerns about the risks and saying the pace at which AI capabilities improve must slow down.

Those events form the backdrop to Newsom's argument that the industry is moving faster than the safeguards meant to govern it, and that rules should be written now rather than after something goes wrong.

A 2028 subplot

Newsom is widely viewed as a possible 2028 Democratic presidential candidate, and his order makes him the latest potential contender to wade into the AI regulation debate as public concern rises.

He is not alone. Pennsylvania governor Josh Shapiro called this week for greater federal regulation of AI while criticizing President Donald Trump for minimizing concerns about it, saying the warning signs cannot be ignored and the risks are real.

Former vice president Kamala Harris, the party's unsuccessful 2024 nominee who is weighing another potential 2028 run, wrote a lengthy social media post saying the frontier of artificial intelligence is advancing at an alarming speed and that a slowdown is critical to ensure AI serves the public interest.

The flurry of statements suggests AI safety is becoming a recurring campaign topic well ahead of the midterms, with Democratic figures competing to define the terms of the debate.

The roadblock in Washington

Regulating AI at the state level could prove difficult. The president signed an executive order in December 2025 blocking states from enforcing their own regulations around artificial intelligence, setting up a direct conflict with any California framework that emerges from Newsom's committee.

If California moves to enforce rules the administration has preempted, the fight is likely to land in court. The Newsom AI regulation framework faces an immediate test in the December 2025 White House order blocking state AI rules. For now, Newsom's panel will focus on drafting the guidelines, with the legal battle likely to come later.

Newsom's order sidesteps the legislature for now, putting the drafting power in the hands of the expert committee and his office.

According to CNN's reporting, the two-month timeline means the state will see a concrete proposal by mid-November, setting up a collision between California's ambitions and federal preemption that could define the AI regulation debate going forward.