California explores AI kill switch but experts question its feasibility

California Gov. Gavin Newsom issued an executive order on Sept. 18 directing an expert panel to explore an "AI kill switch" and strengthen state AI safety laws, including third-party evaluations and mandatory risk reporting. The concept of a universal off button for rogue AI remains undefined, with proposals ranging from a federal bill giving Homeland Security shutdown authority to industry-led safety audits. Experts question its feasibility, noting AI systems aren't contained to single environments and may require global coordination to halt. Critics argue kill switches are "binary" regulations easily outdated or gamed by companies, while developers face financial disincentives to activate them. The debate intensifies amid doomsday warnings, though some AI leaders doubt such mechanisms offer a long-term solution.