America’s Only AI Protections Are About to Get Bulldozed
A leaked Trump executive order would ban state AI laws, punish those that enforce them, and open the door to federal surveillance that states currently fight. The tech industry loves it. Voters don't.
Hello! I’m Jacob Ward, a veteran technology journalist. I’ve been a correspondent for NBC News, Al Jazeera, CNN, and PBS, and I’m the former editor-in-chief of Popular Science magazine. My book The Loop: How A.I. is Creating a World without Choices and How to Fight Back, combines the last 60 years of behavioral science with my experience covering big tech to predict the rise of commercial A.I. and its effects on us. I spent nearly a decade on it and honestly I thought I was years ahead of the topic. ChatGPT shipped 10 months later. I typically spend my time at my own Substack and podcast The Rip Current, but I’m honored to be Hard Reset’s newest contributor interrogating tech’s impact on society and labor. Here’s my first post…
Americans could soon be on an A.I. superhighway with no speed limits, no lane dividers, and no guardrails, if President Trump gets his way. The White House is, according to a leaked draft obtained by several outlets, about to issue an executive order banning state AI laws.
In some parts of the US, A.I. is already observing some speed limits. Six million Americans currently enjoy protection against A.I. systems that might accidentally or unfairly cost them a shot at a house, a job, or health insurance. Another 12 million don’t have to worry that the robotic systems reviewing their employment applications might discriminate against them. And 40 million more Americans get to see inside the security and safety protocols of the world’s biggest A.I. companies.
But those are only for the residents of Colorado, Illinois, and California. And that’s because while all 50 states have passed some form of A.I. legislation, no federal law regulates A.I. in this country. (Decades after the invention of social media, we don’t even have a federal law governing data privacy.)
The executive order directs the Attorney General to attack state A.I. laws, and directs other agencies to withhold certain federal funding from those states deemed to have imposed “onerous” restrictions on the technology. It’s a sweeping federal action from an administration that has been consistently political in its attitude toward states rights, taking away state power on some issues, like immigration and family planning, while handing it out freely on others, like education and healthcare.
Many of the threads that run through the executive order — a need to empower the industry to compete with China, a desire to avoid having to comply with a patchwork of state laws — run back to the talking points of the major tech companies, as well as the investment community that surrounds it. And the companies that stand to benefit from a hands-off federal approach are, of course, among the donors to the president’s new ballroom, a list that also, hey, weird, overlaps with the CEOs who attended Trump’s grand reception for Saudi Crown Prince Mohammed Bin Salman this week, where the two discussed pouring billions into technology and A.I. together.
The first time this crushing of state laws was proposed — when Senator Ted Cruz tacked it onto the Big Beautiful Bill Act this summer — it went down in flames, 99-1, in the Senate. Now it’s alive again in congress, with Republicans planning to attach the idea to the National Defense Authorization Act, which gets a vote at the end of December. But it’s not clear that members are any more enthusiastic about the idea this time, not least because it remains an incredibly unpopular position nationally. A September Gallup poll found that 80% of Americans believe in the need to regulate A.I. even if it slows down the development of the technology. Even Florida governor Ron DeSantis took to X to call the EO “executive overreach.”
So why plow ahead? The executive order cites national security, a need to fight against “woke” A.I. regulations in blue states, and the need to help the industry grow rapidly in arguing against state A.I. laws. But there could be other motivations here as well.
For instance, doing away with state laws on A.I. would clear away powerful forces stymieing the administration’s efforts to assemble a surveillance apparatus using the services of for-profit companies like Palantir, as Cody Venzke, Senior Policy Counsel for the ACLU, explained to me.
“It has been shown over and over again that the federal government, especially some of its intelligence bodies, are happy to buy data that they could not otherwise obtain,” he said.
“State regulations cut off that data collection at its source so that your Candy Crush game on your phone is no longer broadcasting your location to a data broker, or it’s no longer available for some intelligence apparatus within DHS to purchase.”
Can the EO survive in court? Brad Carson, president of Americans for Responsible Innovation, described it as an incorrect reading of commerce law, writing on X that it’s “the tobacco strategy: fight all laws and then sue once passed.”
Meanwhile, the states who’ve passed these laws, like California, will almost certainly sue. Scott Wiener, the California State Senator who authored the state’s sweeping AI law, SB-53, referenced in the executive order (and who is running to replace Nancy Pelosi in Congress), told Politico that the President “has no power to issue a royal edict canceling state laws.”
What We’re Reading
More than 150 community members in Michigan gathered to push back against yet another data center coming to Kalkaska County.
Stories about the AI bubble are suddenly everywhere. AI Now had an opinion piece in the WSJ last week entitled “You May Already be Bailing Out the AI Business”. Cade Metz, the NYT’s venerable AI reporter, broke down what’s going on in Silicon Valley and New York for The Daily yesterday. And David Dayen, writing for The American Prospect, looked at the sketchy financial engineering underpinning some of the massive valuations.
Cade Metz also had a story this week about the large numbers of Chinese Nationals working in Silicon Valley AI labs. There’s strong cognitive dissonance there when you put that labor force (and tech companies’ desperate desire to sell in China) against the anti-China fearmongering from so many in the sector.
Verizon announced 13,000 layoffs. Google is offering voluntary buyouts to staff in the UK.
OpenAI is down a board member. Who do you think they’ll choose to replace Larry Summers?
“If AI causes mass unemployment or severe fiscal shocks, elected officials and policymakers will need to act fast to limit the disruption. Fortunately, some of the most reliable and powerful options are also the most familiar.” - Kevin O’Neil, from the Rockefeller Foundation, makes the case for taxing AI.
And because we like to end on a good time, see what happens when tech bros discover a public park for the first time.