California Governor Newsom Signs Executive Order on AI Safety, Revives 'Kill Switch' Proposal
Governor Gavin Newsom has signed an executive order directing California state agencies to develop comprehensive AI safety plans, marking a new phase in the state's approach to artificial intelligence oversight.
The executive order requires state agencies to establish safety protocols for their use of AI systems, building on earlier concerns about potential risks associated with the technology. The directive also revives the concept of a "kill switch" mechanism—an emergency shut-off capability that could be deployed if an AI system behaves unexpectedly or poses risks to public safety.
The move comes after Newsom previously rejected SB 1047, a more comprehensive AI safety bill that would have imposed broader regulations on AI companies operating in California. That veto signaled a preference for a more measured regulatory approach focused on state agency operations rather than industry-wide mandates.
To implement the new framework, the governor is establishing a dedicated panel to examine AI safety regulation and develop specific guidelines for state agencies. The panel will be tasked with evaluating the feasibility of various safety measures, including the proposed kill switch concept.
State officials have emphasized that the executive order applies specifically to government use of AI systems, rather than imposing new requirements on private companies or AI developers. The approach aims to ensure responsible deployment of AI within state operations while the broader regulatory landscape continues to evolve.
The timeline for the panel's recommendations and implementation of agency-specific safety plans remains under development, with officials indicating that stakeholder input will be sought throughout the process.