California pushes ahead with an AI “kill switch” and tougher independent oversight

California Governor Gavin Newsom is moving from vision to implementation on AI oversight, advancing an executive order that calls for an independent watchdog and technical capabilities often described as an “AI kill switch.” The initiative, first outlined in September, is now being translated into concrete steps such as formal audits of how state agencies deploy generative AI and requirements that critical systems can be rapidly paused or isolated when serious incidents are detected. With comprehensive federal AI legislation still stalled, California is positioning itself as a regulatory testbed, signaling that major cloud platforms and model providers will be expected to participate in transparency reporting and third‑party evaluations. Industry groups, however, warn that poorly designed rules could slow innovation or drive AI investment elsewhere, and are lobbying for flexible, risk‑based standards instead of rigid mandates. How the state ultimately writes these rules will be closely watched, as it could set a de facto template for AI safety governance across the United States.