Newsom Goes Easy on AI — for Now

SACRAMENTO, Calif. — The other day I searched for some of my articles regarding the housing issue and, much to my surprise, read an artificial intelligence summary of my position that misstated it entirely. It claimed I am associated with NIMBYs (Not In My Back Yarders), when my positions closely echo YIMBYs (Yes In My Back Yarders). I searched on another AI tool, which captured my position accurately. It’s an emerging technology, so it’s best to verify whatever it is you might read.

That painfully obvious warning aside, AI offers stunning potential in every field of human endeavor. And it’s advancing so rapidly that it’s rather pointless to try to quash it. It’s foolhardy to believe governments, which struggle to efficiently provide age-old services (infrastructure, policing, budgeting) can get ahead of the curve of a technology so advanced that even the people who work in the field don’t entirely understand it. The most important fact to know about Luddites is how unsuccessful they were at stopping mechanized looms.

Yet government officials (in liberal and conservative states, but mostly in the former) aren’t lacking in self-confidence with regard to AI regulation. As my R Street colleague and AI expert Adam Thierer wrote in July, the U.S. Senate’s failure to pass a 10-year moratorium on state-level AI regulations cleared the way for blue-state lawmakers to push forward the Biden administration’s “fear-based” agenda that sought to tamp down on AI development. At the time, lawmakers had introduced more than 1,000 AI-related bills that would do little more than threaten “America’s global competitiveness, especially against China.”

When it comes to tech regulation — or any other type of regulation, actually — California leads the way and usually not in the right direction. Our state’s role here is significant, not just because the AI industry largely is based in the Bay Area – but because California’s marketplace is so large that any state tech rules become de facto national rules. I can’t tell you how many times California lawmakers have proposed some new regulation (cap and trade, for instance) by bragging about the way it will influence the national or international regulatory framework.

Because the U.S. Senate failed to control an avalanche of state bills, the tech industry is busy playing whack-a-mole. Fortunately, it stopped what could have been the most troublesome threat: California Assembly Bill 1064. The legislation would “prohibit a person, partnership, corporation, business entity, or state or local government agency that makes a companion chatbot available to users 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.”

How could any AI business be “foreseeably capable” of preventing its complex tools from “doing certain things” that could potentially cause harm to any child? Enforcing such vague and overly broad language would be a nightmare. The only logical result is that AI developers would have to figure out a way to keep chatbots off the keyboards of anyone under 18. I have many deep disagreements with the Trump administration, but its emphasis on letting the private sector do its thing with a pro-innovation strategy regarding technology is the right approach. It’s not like AI isn’t going to take off — and far better that American companies lead the charge rather than less-accountable Chinese ones.

Despite his penchant for regulation, Gov. Gavin Newsom certainly isn’t about to crush the tech industry as he eyes a presidential bid. So his decision to veto AB 1064 wasn’t unexpected, but it was a relief. Last year, he vetoed Senate Bill 1047, which would have required “a developer, before beginning to initially train a covered model … comply with various requirements, including implementing the capability to promptly enact a full shutdown.” That bill was a convoluted mess.

Unfortunately, he has signed several AI-related bills (including some problematic ones) and dozens 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.

This year he waited until the deadline, but issued an unusually lengthy and detailed veto statement of AB 1064 that touched on the key points:

While I strongly support the author’s goal of establishing necessary safeguards for the safe use of AI by minors, AB 1064 imposes such broad restrictions on the use of conversational AI tools that it may unintentionally lead to a total ban on the use of these products by minors. AI is already shaping the world, and it is imperative that adolescents learn how to safely interact with AI systems. … We cannot prepare our youth for a future where AI is ubiquitous by preventing their use of these tools altogether.

Supporters expressed dismay that Newsom wasn’t protecting the children, but good intentions do not necessarily lead to good legislation. The tech industry ended up supporting — and Newsom signed — Senate Bill 243, which imposes a series of mostly reasonable safeguards on chatbot developers. Unlike others, it does so without obliterating the federal Communications Decency Act provisions that protect online platforms from facing the legal liability of publishers.

Newsom mentioned SB 243 in his AB 1064 message by noting that it “requires operators to disclose to minors that they are interacting with AI … and prevent chatbots from producing sexually explicit material.” As with all regulation, it’s best to start slowly in a way that negates obvious harm rather than impose some massive revamping of an industry. California should have learned that overly broad efforts always have negative, unintended consequences. But I don’t expect that it has learned any such lesson.

So, for now, California and the nation dodged a bullet. But, as Thierer noted, it might be a good time for Congress to disarm these meddlesome state legislatures. And, as usual, Americans need to verify all the information they read, with parents taking a lead role regarding their kids. It’s a fool’s errand to expect the government to do it for us.

Steven Greenhut is Western region director for the R Street Institute. Write to him at sgreenhut@rstreet.org.

READ MORE:

Self-Driving Cars Becoming Unstoppable

Another Transit Shakedown of Taxpayers

Los Angeles Faces an Olympian Task