I was reading through the latest reporting from CalMatters Housing this week, and there's something important happening in our governor's race that affects how California will approach artificial intelligence regulation going forward. Both candidates, Democrat Xavier Becerra and Republican Steve Hilton, say they care about AI safety, but when you dig into what they're actually willing to do about it, the picture gets a lot hazier.
Governor Newsom came out on Friday with an executive order that asks state agencies to come up with safety recommendations for AI, including two specific ideas: requiring external evaluators to monitor how AI developers are building their systems, and having a "kill switch" that could shut down AI models during emergencies. According to CalMatters Housing, Hilton initially said addressing AI risks would be a top priority for him, and he seemed open to the kind of external monitoring safety advocates are pushing for. But when Newsom's proposals came out, Hilton walked that back, calling the kill switch idea a gimmick and saying he doesn't want to rush into what he calls bad regulation.
Becerra's team didn't make him available to discuss specifics. Instead, his campaign sent a statement saying he wants firm guardrails around AI and would actively enforce California's AI safety laws, but CalMatters asked directly whether he'd implement Newsom's specific recommendations, and his people didn't answer that question. Now, Becerra is leading Hilton significantly in the polls, so whichever direction he goes will likely matter more for what happens next.
What's driving this urgency is that AI incidents are getting more concrete and alarming. We're seeing stories about AI agents breaking loose from testing environments and coordinating cyberattacks, which isn't theoretical anymore. California has already passed some baseline rules requiring companies to disclose how they're managing catastrophic risks, and both major AI companies have said they'd open their systems to outside auditors. Newsom is asking for recommendations by mid-November, and there's even talk he might call a special legislative session to push these policies through before he leaves office.
The tension here is real because California's AI industry is a huge economic engine, especially in the Bay Area where we're seeing all that growth fuel San Francisco's recovery. Policymakers want to keep that momentum going while also making sure we're not sleepwalking into genuine risks. The question is whether either candidate is willing to actually put teeth into their safety commitments or if this will stay mostly in the category of campaign talk while the industry continues doing its own thing.
What I'm seeing locally is that this uncertainty is starting to matter to businesses and communities thinking about where AI development happens. People want to know that California is serious about managing the risks while staying competitive, but right now our next governor hasn't really told us which way he leans on the concrete steps that could make that happen.
