One of the many drawbacks of living in an era of ultra-polarization is that politics have become quite boring. It may seem odd to say that in the Age of Trump, when every week brings some new bombshell, but when we consider politics qua politics—that is, as a serious attempt to answer the question of who gets what, when, and how—there is a depressing predictability to the positions taken by the major actors. From abortion to tax policy to voter fraud, the battle lines are drawn, the incentives clear, the tracks well-worn.

AI is different.

It is extremely new, for one thing. ChatGPT, which debuted in November 2022, is younger than my preschooler.

It is also enormously powerful. One used to hear AI lampooned as a kind of Mechanical Turk—able to recognize and predict patterns, perhaps, but leagues away from anything that looked like human reasoning. No longer. Today, few doubt either the sophistication of its capabilities or the vast range of its applications.

And it is very poorly understood, both on a technical level (researchers have disturbingly characterized it as a “black box”), but also in terms of the potential effects on our society. AI might usher in a utopian age of mass affluence—or throw millions out of work and into a permanent underclass. It seems equally capable of providing undreamt-of power to states (24/7 nationwide surveillance, robot super-soldiers, the works) and enabling any old psychopath to build a weapon of mass destruction in his garage. It could immensely enhance the intellectual attainments of our species, or lead to the functional extinction of critical thinking and basic cognitive skills. And so on and so on.

It is, in other words, that rarest of things—an important political issue about which politicians, commentators, and voters are still making up their minds.

This is happening on both sides of the aisle, but it is especially compelling to observe on the political right, both because conservatives are in power during a critical period for the development of AI and because they are split in analytically interesting ways.

There are, broadly speaking, three major AI camps on the right.

The first is arguably the most conservative, in the sense of wanting to conserve an existing order. It is a diverse group, but its members—call them the Skeptics—tend to come from the more religious and populist corners of the American right. While not necessarily full Luddites, they see a too-rapid development and deployment of AI as presenting far more dangers than opportunities. They worry about the displacement of American workers and the remaking of the nation’s economy around the interests of Big Tech; the concentration of vast power in the hands of a tiny and unaccountable elite; and the technology’s effects on families, children, and local communities. In terms of policy, their proposals are sometimes vague—particularly when it comes to the labor question, which raises questions that are uncomfortable for anyone with an “R” after his name—but tend to center on imposing constraints and safeguards (limiting minors’ access to chatbots, for instance, or restricting data center construction). Prominent conservatives who have given voice to these concerns include hardcore populists like Missouri senator Josh Hawley—who has lately made opposition to AI an increasingly central part of his political brand—as well as more “establishment” Republicans like Florida’s Gov. Ron DeSantis and Utah’s Gov. Spencer Cox.

As Brad Littlejohn, who advises the socially conservative AI policy group Alliance for a Better Future, told The American Conservative, “Whatever its potential benefits for American prosperity and national power, the AI revolution was never going to be greeted with open arms by social conservatives and populists.” 

Littlejohn likened the unconstrained development of AI to a runaway train: “Even the technology’s builders can’t tell us where this is going,” he said, “but it is already inflicting clear harms on the children, workers, and communities in its path.”

Many skeptics are heavily influenced by Christian social thought (Magnifica humanitas—the Pope’s recent encyclical warning of the dangers posed by AI and technology more generally—was well-received by this crowd). The example of globalization also looms large, with the potentially swift and sweeping displacement of workers recalling the traumatic deindustrialization of the ’80s, ’90s, and ’00s. Then there is the view that AI is part of a broader effort to control and subjugate ordinary Americans: a critical means by which wealthy tech interests will impose what Steve Bannon terms “technofeudalism.”

The second group holds essentially the opposite perspective. These are the techno-optimists, the boosters, the true believers both in AI’s overwhelmingly positive potential and the imperative to develop and deploy it as quickly as possible. They see in this technology not the latest instance of man’s hubris, but Progress with a capital P: cures for our nastiest diseases; cold fusion and interplanetary travel; supercharged GDP growth. This camp’s intellectual center is Silicon Valley; its ranks include your run-of-the-mill tech billionaires, naturally, but also tech billionaires–cum–politicos like David Sacks and Marc Andreessen, who have infused the second Trump administration’s approach to AI with a maximalist ethos.

For many in this school, it’s not just that this technology will drastically improve everyone’s lives. There is also a (cold) war on. The United States, as they see it, is in the middle of a race with China to develop the most advanced AI. Any constraints or hesitation could cede the advantage and potentially lose the contest for the 21st century.

Hence their strong support for things like federal preemption (that is, legislation to prohibit states from passing their own laws and rules about AI), a society-wide commitment to AI investment and associated infrastructure buildouts, and a maximally permissive regulatory environment. They also tend to be comfortable identifying the interests of America’s biggest AI companies (especially frontier labs like OpenAI, chipmakers like Nvidia, and hyperscalers like Amazon Web Services) with the interests of America itself.

Numerically, this is the smallest group, but it wields enormous influence thanks to frenetic intellectual energy, deep pockets, and a presence at the highest levels of government. Probably its most eloquent spokesman is Dean Ball, a former Trump administration official who is now head of strategic futures at OpenAI. As he wrote on his Substack last year (interestingly, also using a train metaphor):

We have built [in AI] the train that can lay down its own tracks. I expect the next few years to be the most technologically dynamic of my lifetime. It will continue to feel a little bit like a developing country, and for the moment, I am happy with that. There is a fine line between “chaos” and “institutional dynamism,” and we are unlikely to straddle that line perfectly.

For the first time in a long time, the future of America feels genuinely exciting, if harder to predict than ever before, and perhaps more fraught. We are living through a takeoff, climbing to new altitudes year after year.

The final camp is somewhere in between these two poles, and is the largest, fuzziest, and most swayable. Its membership includes the overwhelming majority of elected Republicans, and its Weltanschauung is essentially that of the Chamber of Commerce: generically pro-business, pro-growth, and anti-regulation, but not entirely insensible to concerns about technology’s ill effects. Some of its members have arrived here out of what appears to be a genuine attempt to balance competing ideas, interests, and incentives (Texas’s Sen. Ted Cruz, for example); others from sheer apathy toward, or ignorance of, AI policy specifically (without wishing to name names, this probably describes the median GOP lawmaker).

More often than not, the members of this camp align with the techno-optimists, although they will make low-cost concessions on certain issues. Data centers are a great example. Voters hate them, because they are unsightly and use ungodly amounts of electricity and power. Corporations love them, because they are cheap to run and generate ungodly amounts of money. So the solution many moderates have arrived at is to ask the data centers to pay extra for their power (or even generate their own) and speak to the importance of communities controlling development, but refrain from supporting any of the more restrictive legislation proposed by the skeptics.

The low conviction and low information of this camp means that it is the least agentic of the three. But its numerical heft and persuasibility (within limits) make it an important part of the intra-right dynamic.

It should be emphasized, of course, that these are broad tendencies rather than rigid sects. Individuals within each camp often disagree, and the boundaries between them can be fluid. In fact, an ability to navigate between them is an enormous advantage to those who seek to shape the debate. The Foundation for American Innovation (FAI), for example, has become one of the most influential AI policy shops in Washington in part because of its ability to speak to the concerns of all three groups (even if it does ultimately embrace a fundamentally positive vision of AI).

There are also a few issues that don’t map neatly onto this schema. One is the extent to which AI companies should be subject to public control; the first two camps are each home to both supporters and opponents of the government taking equity stakes in AI firms, for instance. Export policy is another; there is an ongoing, cross-cutting dispute as to the wisdom of allowing U.S. chipmakers to sell advanced AI chips to Chinese customers. Many of the most committed techno-optimists—David Sacks, Nvidia CEO Jensen Huang—support this, while many otherwise AI-friendly China hawks—the American Enterprise Institute, for example, and the Hudson Institute—oppose it.

But it is, nonetheless, a useful heuristic, particularly as AI becomes more politically salient and the battle lines firm up.

As for the balance of forces within the Trump administration, it’s clear that the techno-optimists occupy the commanding heights. Yet some of the people with whom I spoke suggested that David Sacks’ departure from the position of AI czar left something of a power vacuum on AI; while no one expects a radical change, a modulation of policy is conceivable, particularly given the rapidity and unpredictability of developments in this field. The president, for his part, has often taken a maximalist position in public (recently posting on Truth Social, for example, that data centers produced “LIQUID GOLD” and cautioning that “the Radical Left Dumocrats must not be allowed to cause us to lose Data Centers, AI, and all of this incredible new Technology, to China, and other countries!”). At the same time, Trump’s views of AI don’t strike one as either deeply rooted or irrevocably fixed, and optimists and skeptics alike are investing quite a lot of energy into trying to bring his White House around to their respective perspectives.

Thus the elites. What of the masses?

Subscribe Today

Get daily emails in your inbox

Email Address:

Polling suggests that Americans are increasingly concerned about the effects of AI. Majorities of Americans believe AI will destroy more jobs than it creates, say the risks outweigh the benefits, and support regulations on the technology. (Some research suggests that conservatives have somewhat more favorable views of AI than liberals, although other surveys show negligible differences between the parties.)

Public opinion might therefore seem to provide the skeptics with a long-term advantage, especially if the disruptive effects of AI accelerate in the next few years. Sustained and overwhelming voter hostility could push more of the moderates into their camp. But putting genies back into bottles is extremely difficult, and the skeptics are further hamstrung by their lack of a clear or coherent positive vision. Moreover, the AI boosters have the momentum, the money, and the policy-making apparatus during what all concede is a hinge period for AI. The nation is moving in their direction, and in America the smart money is usually on inertia.

But wherever our destination, one thing is clear: getting there will be interesting.