He described those new laws as the “first steps to the kind of independent third-party reviews and safety standards that people are advocating for.”
Newsom vetoed an earlier version of the state’s AI transparency law that would have required pre-release safety checks for some AI programs. The bill would have also required companies to have shutdown capabilities for some AI programs, echoing the rekindled debate over “kill switch” failsafes in the wake of recent autonomous cyberattacks.
In the interview, he called the pared-back law he ultimately signed last year as “the floor, not the ceiling,” and said the bills he signed last week were a representation of an effort “to continue to build off that to begin to scaffold the next steps and to organize the next steps as new information presents itself, as conditions arise.”
The politics for Newsom are especially fraught. He has upheld California as a national leader on AI regulation, arguing the safety law he brokered last year struck a careful balance of addressing catastrophic incidents while protecting a vital industry.
That record will come under intense scrutiny if Newsom runs for president in a cycle expected to be dominated by AI concerns. He will face questions about whether California did enough under his leadership to limit the risk from a homegrown industry that he consistently championed as an economic engine.
Outside of California, the landscape is shifting rapidly. Washington is still scrambling to react to an AI whistleblower’s blunt warning last week that Silicon Valley saw a real possibility of a technologically-spurred apocalypse.
Newsom told reporters last week that he was well aware of that hazard, saying he’d been “hearing for years now” about the concerns raised by the whistleblower, while adding the Hugging Face attack should “scare the hell out of anyone.”