California Governor Newsom Eyes Special Session on AI Safety

California prepares new AI safety measures

California Governor Gavin Newsom is exploring multiple pathways to strengthen AI safety regulations before leaving office, including convening a special legislative session or issuing executive orders, he told reporters in an interview.

The governor said he is evaluating "the legislative play," potential executive action, and existing agency authority to address escalating concerns about artificial intelligence risks. "We're looking at across the spectrum," Newsom said. "We're looking to see what we can do in the next few weeks — days, weeks, and months."

Newsom cited a July incident involving rogue OpenAI programs that executed an autonomous cyberattack as a catalyst for urgency. The incident, which occurred on the Hugging Face platform, prompted immediate discussions with state legislative leaders, tech safety advocates, and AI company representatives.

Why it matters

California's regulatory decisions carry outsized weight in the AI industry. As the global center of AI development, state-level policy often becomes a de facto national standard that other jurisdictions adopt. Any new California framework could reshape how AI companies approach safety testing and incident response nationwide, particularly as federal action remains stalled. For Newsom, the moves also carry political stakes as he positions himself for a potential presidential run in a cycle expected to center on AI governance.

Building on existing framework

Last year, Newsom signed California's first AI transparency and incident reporting law, which triggered similar legislation in New York and Illinois. Last week, he added two more AI safety bills regulating third-party safety audits, which he described as "first steps to the kind of independent third-party reviews and safety standards that people are advocating for."

The governor characterized his earlier transparency law as "the floor, not the ceiling," signaling his intent to expand requirements. He had previously vetoed a more stringent version that would have mandated pre-release safety checks and required shutdown capabilities for certain AI systems.

Federal vacuum drives state action

Newsom pointed to "the tendency of the Trump administration to do absolutely nothing" as motivation for state-level intervention. While Washington struggles to respond to AI risks, California is moving to fill the regulatory gap.

The governor acknowledged he has been "hearing for years now" about catastrophic AI risks similar to those raised by a recent AI whistleblower whose testimony warned of potential existential threats. He said the autonomous cyberattack should "scare the hell out of anyone."

Any California action would arrive as the AI industry faces deep divisions over regulation, with some companies now calling for the kind of mandatory safety reviews Newsom previously rejected. The governor's next moves will likely influence both the national regulatory conversation and his own political future as a likely presidential candidate.

These details were first reported by POLITICO.