Pleasanton, California — Governor Gavin Newsom signed an executive order directing state agencies to accelerate the implementation of independent safety oversight for frontier artificial intelligence and examine technical protocols for an emergency shutoff mechanism, state leaders announced.
The administrative directive instructs the California Government Operations Agency, working in coordination with emergency services officials, to convene international technology experts to develop comprehensive regulatory recommendations within a sixty-day timeframe. Under consideration are standards requiring independent verification organizations to conduct on-site evaluations within frontier development laboratories and verify corporate risk disclosure filings against state safety benchmarks.
The order expedites timelines under recently enacted state laws, including Senate Bill 813, which created certification standards for independent AI auditing organizations, and Assembly Bill 1405, establishing a state registry for technology auditors. State regulators are also tasked with formulating enforceable benchmarks for automated emergency shutoffs, designed to verify that developers maintain fail-safe mechanisms capable of deactivating models during critical system loss-of-control events.
In announcing the directive, Governor Newsom underscored the urgency of establishing robust state-level safeguards, stating that California cannot delay action to implement meaningful accountability standards as advanced computing models develop rapidly.
