We have to do a lot more in the absence of Federal leadership.
That was one of the less colorful comments from Gavin Newsom, Governor of California, the home state to the US tech sector, when asked about the increasingly febrile question of AI regulation, or rather the lack thereof.
He was rather less temperate in his remarks about Donald Trump’s ludicrous attempt to re-brand AI as Super Intelligence in a bid to calm down the calls from within the AI sector itself to slow down the pace of innovation while we all have a little think to take on board some of the safety implications, and maybe come up with some ideas around how not to have the bots kill us all off by 2030 as appears to be the accepted course of events at the moment.
On the decision by the President of the Unites States (sic) to take time out of his packed political and legislative schedule to insist that, as a priority, all US Government documentation past, present, and future be re-written to include the term SI instead of AI, Newsom opined:
What we saw from the White House yesterday was bullshit, and that should scare the hell out of everybody.
To which I say, yes, yes it should.
And I’d go on to say that, in my opinion, suck-ups like Meta’s Mark Zuckerberg ought to feel an acute sense of shame about playing along with this Trumpian appreasement, but I realize that any search for shame in such quarters would be as fruitless an exercise as trying to make the Metaverse a thing. Mind you, that hasn’t stopped on/off Presidential BFF Elon Musk looking to shore up his standing by announcing over the weekend that he’ll be re-naming SpaceXAI as SpaceASI etc.
(In so doing, of course, he’s painting arch-enemy Sam Altman into something of a corner – OpenSI, anyone? And if not, why not, Sam? Didn’t you hear the President’s instructions? Naughty Sam! Let’s hope no-one brings that to Trump’s attention, eh?)
The grown-ups in the room
But back to Newsom and the grown-up back in the Golden State. He argues:
Super intelligence [lower case!] is clearly not coming from the White House – that’s why California continues to lead. AI should expand opportunity – not come at the expense of workers and families. As this technology re-shapes the workplace, California is putting people at the center, ensuring we all have a voice in the decisions shaping their future.
To that end, Newsom’s time in office has seen a raft of AI-related State-level legislation brought on, such as SB 947, SB 951, AB 1331, and AB 1883 signed after Trump’s White House AI sideshow. What outrageous affronts on truth, justice and the American way do these impose? Running down them:
SB 947 – the so called ‘No Robo Bosses Act’ – is a regulation banning employers from relying solely on “automated decision-making systems” to fire or discipline workers and restricts AI’s use as a “principal tool” in these worker decisions. Organizations must have a human reviewer corroborate decisions using additional information, such as managerial evaluations, peer reviews, and personnel files. Additionally, affected employees must be provided with written notice that AI was “primarily used” in their termination or disciplinary decision.
SB 951 – AKA the ‘California Worker Technological Displacement Act’ – is intended to protect workers from the impacts of technological displacement caused by AI. It requires employers to provide a 90-day advanced written notice to affected workers, the Employment Development Department, and state and local entities before any technological displacement affecting 25 or more workers or 25% of the workforce, whichever is less. Employers must also provide a written technology hiring disruption notice when executing a technological reduction in hiring due to the adoption of AI or other automating technology.
AB 1331 is a 2025–2026 bill to restrict employer use of workplace surveillance tools, especially banning monitoring in workplace bathrooms and imposing civil penalties up to $500 per violation.
AB 1883 also prohibits certain workplace surveillance technologies and restricts employer use of employee data, with civil penalties for violations. It specifically targets tech that involves facial recognition, gait recognition, emotion recognition, and neural data collection. Employers are also prohibited from using these tools to infer or monitor protected employee characteristics, such as veteran status, ancestry, religion, disability, or engagement in legally protected activities
All of these have been introduced with the stated aim of protecting the working rights of human beings over the bots rather than specific safety concerns around the tech. California Federation of Labor Unions, AFL-CIO President Lorena Gonzalez says:
Workers across California have demanded that our state lead the way in regulating AI in our workplaces…By signing these four first-in-the-nation bills, Governor Newsom has shown the country how we can begin to put guardrails on AI at work to ensure human oversight over this technology, limit surveillance, and put workers first. We must continue to push for regulations to ensure we control this no technology, not be controlled by it.”
But Newsom and California aren’t just jumping on the AI bandwagon here. It was back in 2023 that the Governor signed the first-in-the-nation Executive Order on generative AI. Since then there has been legislation introduced around:
Frontier AI safety, requiring frontier AI developers to publicly disclose safety frameworks, report specified critical safety incidents to the state, and protect whistleblowers who report serious risks.
Independent AI oversight, making California the first US State with a formal framework for certifying independent verification organizations with sufficient expertise and demonstrated independence from AI companies to objectively assess AI systems and models for safety and risk. A state registry for AI auditors and standards for their independence, transparency and integrity has also been put in place.
Child safety standards for AI companions, introduced via the bi-partisan Adam’s Law. This lays down safeguards around companion chatbots used by children, including suicide and self-harm crisis protocols, parental controls and notifications when safety settings are disabled. California also has laws in place to ban toys containing AI companion chatbots for five years, prohibit the use of student personal information to train AI models, and expand protections against AI-generated child sexual abuse material.
And last month Newsom signed another EO calling on a group of experts to meet and create, within two months, a guide for the state to strengthen its AI safety and security laws. This may result in a requirement for independent third parties to write safety plans for frontier AI companies, or mandating that companies develop a “kill switch” to shut down their models in case of emergency, according to an announcement from Newsom’s office.
But what about China!?!
Now from where I’m sitting as a silly old, bureaucracy-loving European, all of that sounds commendably sensible. From the point of view of the White House, concerns about AI safety are, and I quote, “a hoax”, like COVID or climate change or gun control. Regulations slow down innovation, is the argument.
China won’t be doing regulations, will it? Actually there are a fair few regulations in place around safety there, and anyway, is parity with a totalitarian dictatorship with a track record of repression and human rights abuse really the aspirational level playing field we should be aiming for here?
But no, stifled innovation is too big a price to pay for safety it seems. California is a banana republic in its own right, sneer the naysayers, driving out big business to friendlier climates. Look at the number of tech firms that have uprooted their Golden State HQs and moved away – HPE, Oracle, McAfee Snowflake, X, Tesla, SpaceX (as was), and Palantir among them.
Most have moved to Texas and Florida where there is no personal or corporate income tax, but California’s stricter regulatory regime has been cited as factor in some of these moves.
That said, while there are some impressive scalps in that list, the Public Policy Institute of California has made a point of noting the number of firms whose headquarters have left the state is small, a single figure percentage, and that those who do move HQ out of the states still retain offices and teams in California. The PPIC argues:
The fact is that companies, headquarters, and branches are always entering and leaving states, as well as launching, growing, or shrinking. Businesses must consider many different factors when making these decisions. When it comes to location, California does have high taxes and tough regulations, which could potentially make other states attractive destinations. But the state boasts key advantages as well, including highly educated and innovative workers and population density, both of which can make businesses more productive. The state also offers a good climate and appealing quality-of-life amenities, including its environment, natural beauty, and rich and vibrant culture.
And for every Oracle or HPE or X making an exit, there’s an Apple, a Salesforce, and a (so far) Meta staying put, not to mention the leading frontier model AI firms, such as OpenAI and Anthropic. And some of these firms have been the most vocal about the need for regulation, not just around AI, as Anthropic has recently done, but on topics such as data protection and negligence by social media platform providers.
Salesforce, for example, has long held up California’s State-level work as an exemplar to the Federal Government, whose inability to get its act together on national data protection has been a long-standing political and legislative failure. But then as CEO Marc Benioff noted at the top of this year, talking about social media abuses, long before the current AI brouhaha kicked off:
These US tech companies, they hate regulation…They hate regulation.
In fact, there are signs that some of those firms have seemingly learned a lesson. In 2024, OpenAI’s Chief Strategy Officer Jason Kwon was busy arguing to anyone who’d listen that California’s SB 1047 — the bill that first proposed a kill switch mandate — would “threaten that growth, slow the pace of innovation, and lead California’s world-class engineers and entrepreneurs to leave the state in search of greater opportunity elsewhere.” It spent 2024 and early 2025 spending heavily to lobby against mandatory kill switch requirements.
But by last month, OpenAI was welcoming Newsom’s signed kill switch EO as “an important step toward adaptable national safeguards” and asking for SB 53’s remit to expand to cover the training and evaluation phases, these being the point at which its own rogue models seemingly escaped containment.
California matters
Nonetheless, such hints at Pauline conversion aside, most of the evidence still points to Benioff’s point being well-made. That being so, the consolation such anti-reg firms have is that even if they don’t have an unquestioning ally in the Governor’s Mansion in California, they do in the White House right now. And the man in charge there is no more a fan of California, which he rages is “a crooked state” and a “third world nation”, than he is of regulation of the AI sector.
But with a nominal GDP in 2025 of $4.25 trillion, California would rank as the world’s fourth largest economy in its own right were it to secede from the United States, behind the US, China and Germany, but ahead of the likes of Japan, the UK, India, France and Russia. So what happens in California still matters, despite Trumpian tantrums to the contrary, and that means what the State decides to do in terms of AI regulation sets an important, and welcome, precedent.
Newsom’s term as Governor is up in January and attitiudes towards AI and regulation are likely to feature strongly in the choice of his successor. As for Newsom himself, he’s yet to declare himself a candidate for US President in two years time, although that’s long been assumed to be his next career move.
But even if he doesn’t end up running, let alone making it to the Oval Office, AI safety and regulation policy is now on the mainstream political agenda and will play a major role in the national Presidential campaigns. The current mid-term polling shows that the White House is massively out-of-step with the electorate on this one, even among Republican voters, outside of the hardcore MAGA brood.
Let the people speak as the frontier model leaders fail to stand up for what is right.