AI POLITICS - Becerra and Hilton agree artificial intelligence carries serious risks. What Californians still do not know is precisely what either candidate intends to do about them.
California has spent decades trying to regulate new technologies after they transform society. Artificial intelligence may not give us that luxury. The next governor will take office as AI advances at extraordinary speed, businesses deploy increasingly autonomous systems and policymakers confront questions that only a few years ago sounded like science fiction: What happens when an advanced AI system behaves in ways its developers did not anticipate? Who has the authority to stop it? And who is responsible when it causes real-world harm?
Those questions deserve more than campaign generalities. Yet Californians still do not have detailed answers from either of the two candidates seeking to become their next governor. Democrat Xavier Becerra and Republican Steve Hilton have both acknowledged the importance of AI safety, but neither has committed to the full set of safeguards Gov. Gavin Newsom has directed state agencies to examine, including independent safety evaluations and mechanisms capable of shutting down certain AI systems during emergencies. That is more than another policy disagreement. It is a preview of one of the most consequential responsibilities awaiting California’s next governor.
Agreement on the Risk, Uncertainty on the Response
Hilton has described AI safety as an urgent priority and expressed interest in outside safety evaluations. But he rejected Newsom’s proposed “kill switch,” calling the concept a gimmick, and cautioned against responding to growing concern with poorly designed regulation. That concern deserves serious consideration. Badly written rules can stifle innovation, strengthen established companies at the expense of new competitors and attempt to regulate tomorrow’s technology using yesterday’s assumptions.
But rejecting one proposal raises an obvious question: What would replace it? If an advanced AI system were involved in a major cyberattack or another serious emergency, what authority should California have to intervene? What obligations should developers face before releasing increasingly powerful systems? Hilton should spell that out.
Becerra owes Californians the same specificity. His campaign has said California cannot accept technology that poses catastrophic risks to public safety and that a Becerra administration would establish guardrails around genuine harms while enforcing the state’s AI safety laws. Those are significant principles, but they are not yet a governing blueprint. Would Becerra require independent evaluations of the most powerful AI systems? Would he support an emergency shutdown mechanism? Would he implement the recommendations Newsom has ordered state agencies to produce? Californians should know before Election Day, not after inauguration.
Newsom’s Reversal Shows How Quickly the Debate Is Changing
The uncertainty surrounding AI policy is not limited to the candidates. Newsom himself illustrates how dramatically the political landscape has shifted. In 2024, he vetoed Senate Bill 1047, legislation that would have imposed stronger requirements on major AI developers, including third-party safety auditing and shutdown capabilities. At the time, Newsom warned that the legislation could unnecessarily restrict innovation. Two years later, his administration is examining some of those same concepts.
That evolution demonstrates something important: AI policy cannot be reduced to a simplistic contest between innovation and regulation. California needs innovation, but it also needs credible safeguards for circumstances in which powerful technology could threaten public safety. Those goals are not necessarily incompatible. The harder question is where government draws the line.
California Cannot Pretend This Is Someone Else’s Problem
California has a unique stake in answering that question. The state is home to many of the companies developing the world’s most advanced artificial-intelligence systems. The industry is producing jobs, investment, technological breakthroughs and enormous economic activity. That gives California an extraordinary opportunity, but it also gives California an extraordinary responsibility.
The state has already begun building a regulatory framework. Senate Bill 53, signed by Newsom in 2025, established transparency requirements for large AI developers concerning catastrophic risks and requires certain critical safety incidents to be reported to the state. California has since enacted additional measures establishing frameworks for independent verification and AI auditing.
But transparency and auditing alone do not settle the larger debate. What happens when safeguards fail? California needs a serious public discussion about independent testing, emergency authority, cybersecurity, human oversight, accountability and the responsibilities that accompany the release of increasingly capable AI systems. That debate should not be driven by panic, nor should it be dictated by the technology industry. It should be driven by evidence.
The Questions Both Candidates Should Answer
The gubernatorial campaign provides an opportunity to bring that debate into the open. Becerra should explain what his promised “guardrails” would actually contain. Hilton should explain what safety framework he would establish if he believes Newsom’s emergency shutdown proposal goes too far.
Both candidates should explain what circumstances would justify state intervention, what independent testing they would require, how California should respond to serious loss-of-control incidents and how they would prevent regulation from becoming a barrier that protects today’s dominant technology companies from tomorrow’s competitors. These are difficult questions. That is precisely why candidates for governor should answer them.
California does not need candidates competing to sound the most frightened about artificial intelligence. Nor does it need candidates competing to sound the most enthusiastic about it. It needs specificity. The next governor will inherit a state sitting at the center of one of the most consequential technological transformations of our time.
Artificial intelligence could produce extraordinary advances in medicine, science, education and economic productivity. It could also create risks that existing government institutions were never designed to confront. California’s challenge is to capture the benefits while establishing credible safeguards against the risks.
Becerra and Hilton agree that AI safety matters. Now Californians deserve to know what each of them would actually do about it. Because when technology is moving this quickly, “we’ll figure it out later” is not an AI policy.
(Avrohom Moshe is a California-based civic affairs writer and public policy analyst focused on government accountability, fiscal responsibility, and public safety. He also writes on issues affecting the Jewish community, including combating antisemitism and advancing civic engagement.)
