Sacramento — California officials are adamant about regulating the emerging Artificial Intelligence industry even though most of the world’s top AI companies are based here. A heavy-handed approach toward this burgeoning industry will not only harm the state’s economy, but the national one as well given that California’s huge market makes our regulations the de facto national standard. That’s why it’s refreshing to see the Trump administration, which is hardly a bastion of laissez-faire thinking, take a hands-off stance on this important issue. My point: When it comes to technology, the state has a long track record of failure. As CalMatters reported this week, the Trump administration and congressional Republicans have angered California officials by circulating a draft executive order that attempts to limit the ability of states to pass their own AI rules. The president proposes creating a litigation task force with the sole responsibility of challenging state regulations. “American AI companies must be free to innovate without cumbersome regulation,” the draft EO explained. “But state legislatures have introduced over 1,000 AI bills that threaten to undermine that innovative culture. California, for example, recently enacted a complex and burdensome disclosure and reporting law premised on the purely speculative suspicion that AI might ‘pose significant catastrophic risk.’” That’s a correct read on what’s going on. It’s awfully hard for a company to innovate if it faces an avalanche of proposed rules in 50 legislatures. This is a good sign given that the Republican-controlled Senate last year killed a measure that would have passed a 10-year moratorium on state AI controls. So far, California Gov. Gavin Newsom has taken a moderate course on such bills. Last session, he vetoed the most noxious one: Assembly Bill 1064 would have prohibited companies “from making a companion chatbot available to a child unless the companion chatbot is not foreseeably capable of doing certain things that could harm a child.” Heaven knows how a company or government agency could determine whether an algorithm is “foreseeably capable” of doing “certain things” that could lead to a child’s “harm.” That’s insanely broad and imprecise language that would lead to endless litigation. AI is arguably the most complex innovation the world has ever seen, so it’s mind boggling to suppose that bureaucrats in some Sacramento office, or judges in the civil court system, could parse out the foreseeability of doing potential harm. Talk about a bill based in hubris. But, as I wrote for The American Spectator last month, “Unfortunately, [Newsom] has signed several AI-related bills (including some problematic ones) and dozens of others will surely make their way to his desk next year. In effect, Newsom and the California Legislature are now making AI policy for the entire nation instead of Congress, which is extremely troubling for the future of the industry.” Other states — and not solely liberal ones — will likewise hobble the industry with their own peculiar concerns. The president is rightly worried this will propel China into the forefront of this technology. In a letter to congressional leadership cited by CalMatters, California Attorney General Rob Bonta made what sounds like a reasonable argument not to pre-empt state AI regulations: “Preemption of state AI laws would undermine these and similar efforts across the country to enact common sense guardrails on AI, and deny all Americans the benefits of state-led experimentation and progress in this rapidly evolving space. Any federal AI law should serve as a floor, not a ceiling, preserving flexibility for states to go further where necessary to protect their residents.” Most Americans have no problem with “common sense guardrails” on this powerful technology, especially to protect children from deception and abuse. Furthermore, states are indeed the laboratories of democracy. As a believer in federalism, I usually defer to the states to come up with their own rules, then the rest of the nation can get out the popcorn and watch how it plays out from afar. Sadly, California’s laboratory usually serves as evidence of what the other 49 states should avoid, but such is life in our decentralized system. Bonta’s argument doesn’t work when it comes to a technology that is national or really international in nature. California is essentially doing with AI as it always does with environmental law and labor policy by trying to write laws for the nation, which isn’t real federalism. As my R Street colleague and AI expert Adam Thierer explained in his congressional testimony in September: The Constitution assigns Congress the lead role in protecting interstate commerce. The Founders wisely provided Congress with this power so that it could facilitate commerce between the states by eliminating barriers that states might otherwise be inclined to erect if left to their own devices. It is essential that Congress exercise that responsibility promptly to ensure the robust development of the national AI marketplace. And allow me two slight detours. First, some San Francisco officials have been pushing for the Legislature to let localities determine their own rules to regulate self-driving taxis such as Waymo after one of those robotaxis ran over a beloved neighborhood cat. As I pointed out in my newspaper column, 5.4 million cats are run over each year, with 5,399,999 from the wheels of human drivers. Autonomous Vehicles are indeed a form of AI, and offer myriad life-saving benefits for critters and, most important, for human beings given that they are better drivers than people. California’s lawmakers understand that the state can’t allow every one of its 483 municipalities and 58 counties to impose their own driving rules or bans if we hope to see this technology take off. The Legislature has killed efforts to let the locals dictate state driving policy, so likewise the feds need to tamp down on state efforts to dictate technology policy. Second, a recent Sacramento Bee investigation revealed that California’s state government spent $450 million since 2019 to revamp its antiquated 911 systems and several years later found the system didn’t work. The state is coming up with yet another new system that will take years and many more millions of dollars to implement. My point: When it comes to technology, the state has a long track record of failure. The 911 system is far less complex than AI. Who would trust the state to know what it’s doing regarding AI systems? The feds should go with their instincts, get the government out of the way and let the AI market evolve. If there’s a need for “common sense” rules, then let Congress do it for the whole country, rather than let California and other states constantly move the goal posts in a haphazard manner. READ MORE from Steven Greenhut: LA Is Destroying Its Housing Market California’s Hypocrisy on Property Rights Post Prop. 50, California GOP Needs Reality Check Steven Greenhut is Western region director for the R Street Institute. Write to him at sgreenhut@rstreet.org.