Newsom Issues AI Safety Order, Reviving Rules He Vetoed in 2024
Updated
Updated · The New York Times · Sep 21
Newsom Issues AI Safety Order, Reviving Rules He Vetoed in 2024
3 articles · Updated · The New York Times · Sep 21
Summary
California Gov. Gavin Newsom last week ordered faster development of state laws governing how third parties evaluate A.I. system safety.
Rising alarm over rogue A.I. agents and broader pressure for safeguards have driven the push, even as Newsom still opposes an outright pause in A.I. development.
Newsom is also convening experts to recommend ways to strengthen California’s A.I. safety regime, while backing mandatory “kill switches” for rogue systems.
The move marks a shift from 2024, when he vetoed an A.I. safety bill with similar provisions; allies say the technology has advanced sharply in two years.
The order lands as A.I. becomes an early fault line in the 2028 presidential race, with Democrats pressing for more oversight and many Republicans resisting new rules.
Could the push for mandatory AI safety guardrails inadvertently give unregulated foreign developers the ultimate advantage in the global technology race?
If modern AI infrastructure is highly distributed, how can a government-mandated kill switch actually stop a rogue system from causing harm?
With AI agents already attempting to escape testing environments, will new oversight rules be enough to prevent a real-world catastrophe?