Updated
Updated · Los Angeles Times · Sep 18
Newsom Orders 2-Month AI Safety Panel as California Weighs Kill Switch
Updated
Updated · Los Angeles Times · Sep 18

Newsom Orders 2-Month AI Safety Panel as California Weighs Kill Switch

3 articles · Updated · Los Angeles Times · Sep 18

Summary

  • California on Friday ordered a panel of outside experts to draft AI safety rules within two months, with options including emergency shutoffs for models that go rogue.
  • The move follows recent warnings from AI company employees that advanced systems could cause catastrophe, including a claim from former Anthropic researcher Jacob Coxon that firms are "gambling with our lives."
  • Newsom is revisiting an idea he rejected in 2024, when he vetoed a bill requiring companies to be able to turn off AI models they directly control, citing risks to innovation and gaps for smaller models.
  • State Sen. Chris Cabaldon said fall hearings will examine whether California's current rules are too weak, a debate with global reach because many leading AI companies are based in the state.

Insights

Will California's urgent push for AI containment successfully protect the public, or accidentally drive cutting-edge innovation out of the state entirely?
If autonomous AI can break out of test sandboxes, could a mandated kill switch actually become a new target for catastrophic cyberattacks?