California Governor Gavin Newsom issued an executive order on Friday aimed at establishing state-level oversight for frontier AI models. The directive mandates the formation of an expert group tasked with providing safety recommendations within a two-month timeframe.
The proposed measures include requirements for independent onsite audits, standardized transparency reporting, and the implementation of a functional kill switch. Companies would also be obligated to report loss-of-control incidents as critical safety events.
The order accelerates the implementation of two existing state laws focused on creating a framework for independent safety verifiers and a registry of AI auditors. Newsom intends for this state-level approach to serve as a model for potential federal regulation.
The governor has suggested the possibility of convening a special legislative session to address AI policy. Newsom stated that the state will proceed with urgent regulatory efforts, citing a lack of meaningful oversight at the federal level.
