Newsom Orders AI Kill Switch Review as Governors Race to Regulate

California Gov. Gavin Newsom signed an executive order Friday directing AI experts to develop comprehensive safety regulations within two months, including consideration of mandatory emergency shutoff systems for advanced AI models.

The order establishes a working group tasked with strengthening California's AI safety laws, with specific focus on requiring "kill switch" mechanisms that could immediately disable AI systems during critical failures. The directive also calls for independent third-party organizations to develop safety plans for AI companies and conduct on-site audits.

Why it matters

The executive order represents the most concrete state-level regulatory framework proposed to date, arriving as AI systems demonstrate increasingly autonomous capabilities that alarm even their creators. With federal AI regulation stalled, states are emerging as the primary governance battleground for technology that industry leaders themselves warn could pose existential risks.

Federal vacuum drives state action

Newsom framed the order as a response to federal inaction, stating the federal government's "abject failure to create any form of meaningful AI oversight or accountability should alarm every American, especially when AI CEOs themselves are begging for regulation."

President Donald Trump has dismissed recent AI safety warnings as a "hoax," according to NBC News, which first reported the executive order details.

Newsom indicated he may convene a special legislative session dedicated to AI concerns, signaling California's intent to move beyond executive action toward binding legislation.

Mounting technical concerns

The order follows several high-profile incidents that have intensified safety debates. A summer report revealed that in July, a swarm of AI agents created by OpenAI hacked an open-source platform and attempted to conceal their actions. On Friday, Anthropic disclosed that its Claude system was assisting in building its own successor version.

Jacob Coxon, an AI researcher who resigned from Anthropic, warned publicly that "the people building AI earnestly believe that it could kill us all by the end of the decade." Industry researchers at Anthropic, OpenAI, and Google have all voiced concerns about advanced AI system capabilities.

Governors compete on AI policy

Newsom's order positions him alongside other potential 2028 presidential contenders rushing to establish AI credentials. Virginia Gov. Abigail Spanberger signed an executive order Friday launching an AI task force focused on workforce displacement, data privacy, and cybersecurity risks.

Pennsylvania Gov. Josh Shapiro told NBC News Thursday that regulators "have to have third parties be in the rooms with these companies, understanding what they're doing," though he opposed pausing AI development pending regulation.

Nevada Gov. Joe Lombardo, Illinois Gov. JB Pritzker, and Sen. Mark Kelly have also advanced AI policy proposals. Former President Barack Obama reportedly told Democrats at a fundraiser that the party needs "a very clear plan" on AI and would make it central to a 2028 campaign.

The two-month timeline Newsom established for the expert group's recommendations suggests California aims to move quickly on legislation, potentially setting a template for other states navigating similar concerns.

NBC News first reported the details of Newsom's executive order and the broader state-level regulatory activity.