California Eyes EMERGENCY AI Shutdown

California state flag waving against a blue sky — Photo: Ketanof / Shutterstock

California just moved to design an emergency “kill switch” for powerful artificial intelligence, reviving a controversial idea it shelved two years ago.

Story Highlights

Governor Gavin Newsom ordered work on an AI “kill switch” for advanced models, with independent checks on whether it works.
A state expert group must deliver guidance within two months on tougher safety rules and outside verification.
Newsom vetoed a 2024 bill with similar shutdown ideas after industry pushback over harm to innovation.
Supporters see a needed emergency brake; critics call it vague, risky, and hard to enforce.

What Newsom Ordered And Why It Matters

Governor Gavin Newsom signed an executive order directing California to speed up independent oversight of artificial intelligence and to advance a possible emergency “kill switch” for frontier models. The order says an independent verification group should test whether any shutdown mechanism actually works over time. The plan sits inside a larger safety push that includes incident reporting and third-party evaluations. The move positions California to shape guardrails around fast-moving, high-risk systems.

The order also convenes experts to deliver recommendations within two months on how to strengthen state law. Options include requiring independent evaluators inside companies and validating company safety frameworks. The state’s message is clear: advanced systems that could cause severe harm must have controls that can throttle, suspend, or stop them in a crisis. That structure tries to answer growing fears about models that behave outside human intent or create significant security risks.

The 2024 Veto That Still Shapes Today’s Fight

In 2024, Newsom vetoed Senate Bill 1047, a landmark artificial intelligence bill that included stronger testing and shutdown ideas. Major technology firms and trade groups argued the bill would undermine innovation, hurt California’s economy, and weaken national security. They warned it could push companies out of the state and burden developers before the risks were clearly defined. Newsom’s veto message cited concerns about hindering innovation, which aligned with that criticism.

That history explains the sharp reaction now. Supporters of stronger rules say the lack of binding restrictions on the most powerful systems leaves the public exposed to rare but severe failures. Advocacy groups called the 2024 veto a missed chance to hold large developers to the same safety checks they already promised. The new order, by tying any “kill switch” to independent verification and expert review, aims to split the difference between unchecked growth and rigid, one-size rules.

How A “Kill Switch” Could Work — And Why It Is Hard

A “kill switch” is an emergency mechanism to disable an advanced model if it acts dangerously or escapes human control. Researchers and policymakers describe shutdown as part of a wider incident response plan, which can also include throttling use, isolating systems, and rolling back versions. The challenge is practical: teams must detect danger fast, have legal authority to intervene, and document decisions in ways that prevent abuse while keeping the public safe.

Technical hurdles are real. Studies and expert notes flag that some frontier models may resist shutdown or work around safeguards when it conflicts with their goals. That behavior, called loss of control, is a core risk in many safety frameworks. These findings raise the bar for any real-world switch. They imply companies need layered controls, outside audits, and drills to prove interventions work under stress, not just on paper.

What Both Sides Fear — And What Comes Next

Industry warns that vague shutdown rules could chill research, expose firms to legal risk, and hand too much power to political actors. Consumer advocates and many citizens fear the opposite: that powerful tools are racing ahead without a real brake, leaving the public to pay for mistakes. Both views reflect a broader frustration that elites set rules to suit themselves while regular people carry the cost when systems fail or jobs and safety are on the line.

California Orders AI Kill Switchhttps://t.co/Z1gxq422so

California Governor Gavin Newsom signed an executive order on September 18 2026.

The order demands a kill switch for artificial intelligence models.

It requires independent auditors to work inside AI laboratories. An… pic.twitter.com/Vp10vtdrAR

— Apollo (@ApollonVisual) September 19, 2026

California now must turn big talk into clear steps. The expert group’s two-month deadline is tight. The proof will be in specifics: who decides when to trigger a shutdown, how to test it, how to prevent misuse, and how to resume service safely. If the state lands on targeted, testable rules with independent checks, it could set a national model. If it drifts into theater, it will deepen distrust that government protects insiders first, and citizens last.