In October of last year, President Joe Biden issued an executive order on AI that made a bunch of lofty but vague commitments, like establishing guidelines for watermarking AI-generated content and protecting people against AI-enabled fraud and discrimination.
Presidential contender Donald Trump barely knows how to use a computer, so it’s unlikely he has any deep thoughts about AI. But the people around him have revealed a chaotic plan for the tech: tear down Biden’s mild protections and replace them with a buffoonish grabbag of idiosyncratic and sometimes contradictory schemes, motivated by a mixture of obscure grievances and the economic interests of Trump’s wealthy donors.
Take the Republican Party’s official policy platform, which is written with Trump’s signature mix of bluster and capitalizing words for emphasis.
“We will repeal Joe Biden’s dangerous Executive Order that hinders AI Innovation, and imposes Radical Leftwing ideas on the development of this technology,” it bellows. “In its place, Republicans support AI Development rooted in Free Speech and Human Flourishing.”
How would that happen? There would be no way to tell from the campaign’s official policy document, which makes no further mention of the tech after those wild promises.
More clues can be found in The Heritage Foundation’s Project 2025, a doorstopper of a document drafted by political operatives close to Trump that calls for sweeping conservative changes to the government if he wins the election. But rather than offering a coherent viewpoint on AI, Project 2025 instead paints a picture that’s familiar to anyone who watched Trump’s first term: an incoherent broth of bluster, cruelty, and xenophobia.
The document miscellaneously calls for more AI in the government, less AI in the government, using AI to spy on Medicare recipients, using AI to supercharge spy agencies, and various broadsides against China. The closer you look, the less any of it seems to have any real depth or connection to Trump’s actual beliefs — not to mention any semblance of comprehensive or even reasonably actionable policies.
For instance, Project 2025 takes the wildly sweeping position that the White House will “stop US entities from directly or indirectly contributing to China’s malign AI goals.” But Trump earlier this year reversed his stance on lawmakers’ proposed TikTok ban, now saying that he no longer supports blocking the video-sharing app in the US. TikTok is owned by Bytedance, the Chinese tech behemoth that also reigns as one of the country’s most prominent AI firms — a position that would be in head-on conflict with the notion that a conservative government would or even could stop “US entities” from contributing to Chinese AI efforts.
That’s not the only place the document’s AI outlook is soaked in hypocrisy. It also accuses China of using AI to establish “authoritarian control domestically and export its authoritarian governance model overseas” — but simultaneously calls for the US to use AI to expand its own spying efforts, saying that the government should employ AI as a way to “better exploit” messy piles of info and classified intelligence data.
As has been widely reported, Project 2025’s main throughline is an overarching push to consolidate power within the executive branch while overhauling the non-partisan government bureaucracy to install loyalists; one of the document’s many recommendations for how to achieve that end is to install partisan loyalists in leadership positions within the intelligence branch. Mix in AI-strengthened surveillance and spying efforts, and it sounds like it’s trying to turn the US into China.
Project 2025’s implications about AI are as horrifying as what it says outright. For example, the document rails against what it cruelly refers to as “abortion tourism,” or women from abortion-restrictive states traveling to blue states to receive reproductive healthcare, alleging that the Center for Disease Control’s (CDC) system of “abortion surveillance” is “woefully inadequate.” To remedy these perceived inadequacies, it recommends that the CDC should collect much more data about women’s reproductive care, and that the federal government should restrict Medicaid funding to states that fail to cough it up. In this fringe, Christian Nationalist fantasy, what’s to stop the federal government from employing AI to help it sift through data to identify and punish suspected abortions?
In a similar vein, Project 2025 is also deeply obsessed with Medicare and Medicaid data, not only calling for an expansion of medical information-gathering but flippantly proposing that AI should be used to “reduce waste, fraud, and abuse” in Medicare. Medicare fraud is a real problem, but so is algorithmic bias — a grim reality of AI that can and has led to harm against minority groups. The document, though, fails to so much as mention this well-documented flaw in AI tech, and is instead poised to let it loose to make decisions in systems as crucial as healthcare and medicine.
Project 2025 is also intensely focused on clearing any existing supervisory red tape for AI companies. That seems part rooted in its abiding quest to beat China to AI dominance, and part in the same zealotry for deregulatory Silicon Valley libertarianism that coats Trump’s approach to both AI and crypto. In one breath, it argues that the US Patent Office should roll out the red carpet for AI firms by streamlining and softening its procedures. In the next, it urges that the government should abstain from making expensive AI models in-house, instead turning to private ventures.
If there’s one coherent theme in Project 2025’s mess of an AI policy, it’s that AI should be used to amplify the powers of intelligence agencies by hiring tech contractors. That’s wildly convenient for Peter Thiel, the libertarian billionaire who has very close — and very expensive — ties to Trump’s running mate, Ohio Senator JD Vance. Thiel is the co-founder of Palantir, the secretive data and AI firm that counted the Central Intelligence Agency, National Security Agency, and Federal Bureau of Investigation as early investors. Thiel was labeled America’s “shadow president” amid Trump’s presidency, during which time Palantir’s tech was reportedly used by the administration to do things like track and quash political dissent.
More recently, per Bloomberg, Palantir partnered with fellow government contracting colossus Microsoft “in a bid to sell software, including OpenAI’s GPT-4, to US defense and intelligence agencies for top-secret tasks.” Thiel, it’s worth noting, is just one of many influential Silicon Valley operators who have endorsed Trump in recent weeks.
The Heritage Foundation isn’t the only influential Trump ally plotting a new course for AI. According to The Washington Post, yet another group of Trump allies are drafting a new executive order that would clear the way for the Trump White House to enact a series of military and intelligence-focused AI “Manhattan Projects” while rolling back “burdensome regulations for companies.”
Considering that the AI industry is currently mostly unregulated, this blinders-on focus on supposed red tape borders on absurdist. In the real world, tech regulations don’t exist to annoy CEOs and slow industry; they’re meant to ensure privacy and security protections for US citizens, establish corporate and government transparency standards, and make sure that safety remains central to technological advancement.
But none of these concerns — transparency, safety, privacy, security — are mentioned in any of these documents’ AI proposals. They also don’t bring up issues of AI companies and copyright, building standards around AI training data, or working towards fairness in algorithms.
Perhaps most strikingly, since Trump and his ilk are obsessed with the specter of immigrants stealing jobs, nowhere do any of the documents address AI replacing people’s jobs.
According to Roxana Muenster, Brookings COMPASS fellow and PhD candidate in Cornell’s Communication department, these glaring holes speak to the sheer lack of comprehension of the scattershot policy plan.
“What was really astounding — I mean, it’s over 900 pages long, and is incredibly dense in some senses. And then you get to one of the most important topics, if not the most important topic, in tech policy, communication policy, and beyond that… so many jobs will be affected by this. And there is very little on it,” Muenster, who addressed Project 2025’s lacking AI policy in a deep dive for Brookings, told Futurism. “There aren’t any detailed plans.”
“Of course, everyone’s kind of playing catch up [to AI policy], because the technology is faster — that’s normal,” she added. But “there’s nothing on training, there’s nothing on privacy, there’s nothing on security and equity, none of that,” she continued, adding that the prospect of repealing the Biden order without a comprehensive AI plan in place is “worrisome.”
“It would just leave a huge vacuum in terms of AI policymaking,” said Muenster, “where people are less safe as a result.”
Instead, Project 2025 and other conservative policy documents heading into the 2024 election center their haphazard approach to AI on a tunnel-visioned focus on the need to beat China to the AI punch without consideration of American civil liberties, altogether painting a disorienting, contradictory, and often broken framework for AI use and advancement rooted in fringe, dystopian isolationist-meets-technocratic ideologies. But that perhaps is because this movement doesn’t see the AI industry as a necessarily governable area of society — but instead, as a tool to mold around larger, sweeping ideological aims.
Project 2o25 is “not intended to be the best version” of AI policy, said Muenster. “It’s intended to be an ideologically beneficial version.”
“Their tech policy really just feels like, ‘let’s tailor this in a way that we can better achieve these other goals that we have,'” she added.
Is any of this a credible threat? It’s hard to say. Trump generates an endless parade of headlines over his scandals and outbursts, but his skills at enacting policy during his first term were uneven. The people crafting these fantasies of surveillance capitalism might genuinely want them, but they’ll need the cooperation of Trump to do so, and that’s never guaranteed, even to his allies.
For a real vision of the role AI would really play in a second Trump White House, the safest bet might be to look at what Trump himself has been using the tech for.
Earlier this month, the former president falsely accused his rival in the 2024 election, current vice president Kamala Harris, of using AI to fabricate images of a crowd greeting her on the airport tarmac in Michigan. Trump went as far as to say that the “creation of a fake image” should be treated as “ELECTION INTERFERENCE,” and that Harris should be “disqualified” — but the images were real, not AI-generated.
Barely a week later, Trump himself took to Truth Social to share what was a fake AI-generated image of Harris wearing red and speaking to an army of Soviet-esque figures, a Communist flag flying overhead; later in the same day, he shared AI-generated images of Taylor Swift and claimed Swift fans — known as Swifties — were endorsing him. He captioned the collection of photos with an enthusiastic “I accept,” and didn’t disclose whether he knew any of the imagery was fake or not. When asked about the phony Swiftie images, Trump told reporters that he didn’t “know anything about them.” The former president has also shared many AI depictions of himself, including fake imagery of him riding lions and praying.
Elsewhere, he told YouTuber Logan Paul that he believes AI could heighten the risk of nuclear war.
“It’s really powerful stuff AI,” Trump told the internet personality, “so let’s see how it all works out.”
Trump’s personal approach to AI, in other words, has been chaos. Not only has the former president reposted AI imagery that he likes, but he’s baselessly thrown around accusations of AI use when he sees real photos that he doesn’t like. There’s no real throughline to follow — just personal convenience.
But chaos has long been a defining cornerstone of Trump’s political career, and it would appear that AI is no exception. Coupled with the conservative push to deregulate the AI industry, and the contradictory, isolationist AI framework put forth by Project 2025 — well, “let’s see how it all works out” seems to capture this conservative movement’s AI agenda in a nutshell.
More on AI and the 2024 Election: Leading AI Chatbots Stumped When Asked About Biden’s Decision to Drop Out
