The artificial intelligence law that Gov. Gavin Newsom of California signed Monday could set a precedent for what Democrat-led AI regulation looks like nationwide.
The first-of-its-kind law directs AI labs to publicly disclose their safety protocols and notify California of any “catastrophic risks” found in the development of advanced AI models. The bill would also create additional protections for tech industry whistleblowers and a network of public data centers, called CalCompute, which provide California-based startups with some AI development capacity.
California’s Transparency in Frontier Artificial Intelligence Act, also known as Senate Bill 53, managed to win support from both AI regulation advocates and certain parts of the AI industry.
For Newsom, the legislative win bolsters his AI policy credentials as he mulls a potential 2028 presidential run amid an environment where Republicans, led by President Donald Trump, are positioning themselves as AI industry standard-bearers.
“SB53 takes a step towards providing more transparency of these AI models, and that’s a good thing,” Rep. Ted Lieu, who last year led a bipartisan effort to draft recommendations for federal AI legislation, told NOTUS. “Gov. Newsom has demonstrated fantastic leadership in this issue.”
Last year, Newsom vetoed a less AI friendly version of the bill after industry backlash. In a statement, Newsom said the bill he signed this year “strikes that balance,” between community safety and AI industry growth. Both versions of the bill were sponsored by state Sen. Scott Wiener.
“AI is the new frontier in innovation, and California is not only here for it but stands strong as a national leader,” Wiener said, adding that the legislation “builds public trust as this emerging technology rapidly evolves.”
Will Rinehart, a senior fellow at the center-right American Enterprise Institute, says the California law might become a “minimum viable product” for what a consensus version of AI regulation can look like between the pro-industry and pro-regulation wings of the Democratic Party.
“What this shows is that there is at least some sort of soft agreement on what can happen on AI safety,” Rinehart told NOTUS. “This is potentially the maximum piece of legislation that could go through the agreement process right now.”
Any form of state legislation regulating the tech industry in California has an outsize influence on the AI industry.
Leading AI labs within companies such as OpenAI, Meta, Google and Anthropic are headquartered in the state. Over the years, they’ve developed deep relationships with California politicians.
AI is a rapidly developing technology that’s increasingly affecting the way Americans find information, communicate and work. If recent improvements in AI suggest what’s to come, Democratic presidential primary candidates will likely face scrutiny of their AI-related platform and policy proposals.
Rinehart says that a legislative framework with support from AI labs such as Anthropic, which endorsed the bill last month, could become an asset for Newsom, particularly if he pursues the presidency.
“He wants to show that he’s trying to be a reformer, and he’s trying to be pro-growth, while at the same time being responsive to his constituents on these other issues,” he said.
Even Rep. Zoe Lofgren, a staunch critic of last year’s AI bill, said she finds this bill “less than perfect” but better than more AI-skeptical alternatives.
“A lot of people in the industry, while they didn’t think it was a perfect fit, they think it’s a substantial improvement above what New York has done, which was over the top,” Lofgren told NOTUS, referring to a stricter AI bill approved by the New York Legislature earlier this year.
When asked about how this bill positions Newsom for a presidential campaign, Lofgren said, “I wouldn’t consider him an expert, let me just put it that way.”
Rep. Nancy Pelosi, the former speaker of the House and another critic of last year’s AI bill, said the California law “came to a place that wasn’t exactly to where we wanted but it was moved much closer to what we wanted.” She did not elaborate.
Other players in the AI industry have also signaled some begrudging support for the law.
Meta, which has spent hundreds of millions of dollars on hiring AI researchers, and billions on AI infrastructure, in a statement to NOTUS called the law “a positive step” in the direction of “balanced” AI regulation.
Collin McCune, government affairs director at the venture capital firm Andreessen Horowitz, said on X that the law “includes some thoughtful provisions that account for the distinct needs of startups,” referring to CalCompute.
The California law will almost certainly attract federal scrutiny, since Trump directed his Cabinet to identify state laws that might obstruct innovation. Trump and Newsom are also arch political enemies, with Newsom of late making headlines for his incessant social media mockery of Trump, who in turn has panned the California governor as “incompetent” and dubbed him “Newscum.”
But Neil Chilson, a researcher at the right-leaning political think tank the Abundance Institute, doesn’t believe the California bill will necessarily obstruct innovation in the short term.
“The biggest of the frontier developers, they’re under the impression that the documents they already have are largely what is required,” Chilson told NOTUS, referring to AI development safety research.
In the long term, however, it’ll be hard to know how the bill “cuts politically,” Chilson said.
“(By 2028,) What does California’s business environment look like?” Chilson said. “Are normal people in California getting the benefits of these tools, compared to, say, people in Texas or Florida?”
But for now, among Democrats, the California bill is a natural consequence of the lack of federal social media regulation that some Democrats have linked to political polarization, mental illness and misinformation.
“We let misinformation and algorithms dictate how young people interact on the web,” Mike Trujillo, a California-based Democratic strategist, told NOTUS. “That lesson is now being learned.”
He added: “The lesson here is, don’t make the same mistake we’d made with social media with AI,” Trujillo added.
Some California lawmakers, who have long been calling for regulation of the tech industry, agreed.
“This has happened multiple times in technology in California; they promise to be good and they never are. Creators by themselves are fine, but when the VC guys come in, it’s all about return on investment all the time — damn the consequences,” Rep. Mark DeSaulnier, whose congressional district stretched between Oakland and Brentwood, California, told NOTUS.
Newsom is “one of the most prominent candidates already,” DeSaulnier said. “Being aggressive is important right now. People want us to fight.”