In 2020, when Joe Biden won the White House, generative AI still seemed like a useless toy, not a new technology that would change the world. The first major AI image generator, DALL-E, would not be released until January 2021, and it certainly wouldn’t put any artist out of business as it was still struggling generating basic images. The launch of ChatGPT, which popularized AI overnight, was still more than two years away. Google search results based on artificial intelligence, which, whether we like it or not, are now inevitable, would have seemed unimaginable.

In the world of AI, four years is a lifetime. That’s one of the things that makes AI policy and regulation so difficult. The wheels of politics tend to work slowly. And every four to eight years, they move in reverse, when a new administration comes to power with different priorities.

That works tolerably, for example, for our regulation of food and drugs, or other areas where change is slow and there is more or less bipartisan consensus on policy. But in regulating a technology that is basically too young to go to kindergarten, policymakers face a difficult challenge. And this is even more true when we experience a sharp shift in who those policymakers are, as will happen in the United States after Donald Trump’s victory in Tuesday’s presidential election.

This week I reached out to people to ask: What will AI policy be like under the Trump administration? Their guesses were all over the place, but the big picture is this: Unlike so many other issues, Washington has not yet fully polarized on the issue of AI.

Trump’s supporters include members of the accelerationist tech right, led by venture capitalist Marc Andreessen, who fiercely oppose regulation of an exciting new industry.

But next to Trump is Elon Musk, who supported California SB 1047 to regulate AI, and has been concerned for a long time that AI will bring about the end of the human race (a position that is easy to dismiss as classic Musk madness, but which is actually quite common).

Trump’s first administration was chaotic, featuring the rise and fall of several chiefs of staff and senior advisers. Very few of the people who were close to him at the beginning of his tenure were still there at the bitter end. The direction of AI policy in his second term may depend on who is watching at crucial moments.

The new administration’s position on AI

In 2023, the Biden administration issued a executive order on AIwhich, while generally modest, marked an initial effort by the government to take AI risk seriously. Trump’s campaign platform says that executive order “hinders innovation in AI and imposes radical left-wing ideas on the development of this technology,” and has promised to repeal it.

“Biden’s executive order on AI is likely to be repealed on day one,” Samuel Hammond, senior economist at the Foundation for American Innovation, told me, although he added that “what will replace it is uncertain.” He AI Safety Institute created under Biden, Hammond noted, has “broad bipartisan support,” although it will be Congress’s responsibility to authorize and adequately fund itsomething they can and should do this winter.

There are reportedly drafts in Trump’s orbit of a proposed replacement executive order that create a “Manhattan Project” for military AI and create industry-led agencies for model evaluation and safety.

Beyond that, however, it’s hard to guess what will happen because the coalition that brought Trump to power is, in fact, very divided on AI.

“Trump’s approach to AI policy will offer a window into tensions on the right,” Hammond said. “There are people like Marc Andreessen who want to step on the accelerator, and people like Tucker Carlson who are worried that technology is already moving too fast. JD Vance is a pragmatist on these issues and sees AI and cryptocurrencies as an opportunity to break the Big Tech monopoly. Elon Musk wants to accelerate technology in general while taking the existential risks of AI seriously. “Everyone is united against ‘woke’ AI, but their positive agenda for how to manage AI risks in the real world is less clear.”

Trump himself He hasn’t commented much on AI, but when he has, as he did in a Interview with Logan Paul earlier this year – seemed familiar with both the prospect of “accelerating defense against China” and experts’ fears of catastrophe. “We have to be at the forefront,” he said. “It’s going to happen. And if that is going to happen, we have to get the upper hand on China.”

As to whether an AI will be developed that acts independently and takes control, he said: “You know, there are people who say it takes over the human race. It’s a really powerful thing, AI. So let’s see how it all works.”

In one sense, it’s an incredibly absurd attitude about the literal possibility of the end of the human race (you don’t get to see how an existential threat “works”), but in another sense, Trump is actually taking a pretty big picture attitude here. .

Many AI experts think that the possibility of AI taking over the human race is realistic and could happen in the coming decades, and they also think that we still don’t know enough about the nature of that risk to formulate effective policies. around him. So, implicitly, a lot of people have the policy of “it could kill us all, who knows?” I guess we’ll see what happens,” and Trump, as he so often proves to be, is unusual especially for simply coming out and saying it.

We cannot afford polarization. Can we avoid it?

There has been a lot of back and forth over AI, with Republicans calling concerns about equity and bias “woke” nonsensebut as Hammond noted, there’s also plenty of bipartisan consensus. No one in Congress wants to see the United States fall behind militarily or strangle a promising new technology in its cradle. And no one wants extremely dangerous weapons to be developed without the oversight of random tech companies.

Meta’s chief AI scientist, Yann LeCun, who is a outspoken Trump criticis also a outspoken critic of AI safety concerns. Musk supported California’s AI regulation bill, which was bipartisan and vetoed by a Democratic governor, and of course, Musk also enthusiastically endorsed Trump for president. Right now, it’s difficult to translate concerns about extremely powerful AI into the political spectrum.

But that’s actually a good thing and it would be catastrophic if that changed. With rapidly developing technology, Congress needs to be able to flexibly formulate policies and empower an agency to carry them out. Partisanship makes that nearly impossible.

More than any specific issue on the agenda, the best sign about the Trump administration’s AI policy will be whether it continues to be bipartisan and focused on the things that all Americans, Democrat or Republican, agree on, such as what We don’t want to. Everyone dies at the hands of a super-intelligent AI. And the worst sign would be if the complex policy questions raised by AI were rounded down to a general view that “regulation is bad” or “the military is good,” which ignores the details.

Hammond, for his part, was optimistic that the administration is taking AI seriously. “They’re thinking about the right questions at the object level, like the national security implications of AGI being a few years away,” he said. Whether that will lead them to adopt the right policies remains to be seen, but it would also have been very uncertain in a Harris administration.