An America First agenda for AI?

On his first trip to Europe as vice president, J.D. Vance lost no time in establishing himself as a force to be reckoned with in global affairs, first at an AI Action Summit in Paris and then at the Munich Security Conference. Although controversy over Europe’s policies on migration and free speech may have made the biggest headlines, the AI conference may prove more significant over the long haul. In the speech before global leaders and tech executives, Vance made a bold bid to secure American leadership in the global AI arms race against China, and rebuked EU leaders for their heavy-handed regulation of U.S. tech companies.

On the face of it, Vance’s speech pursued precisely the two ends one might expect any “America First” statesman to advance abroad. First, it insisted “this administration will ensure that American AI technology continues to be the gold standard worldwide.” Translation: China’s DeepSeek might be impressive, but Europe would be very unwise to dabble in tech partnerships with China, which hardly has their best interests at heart. Second, Vance argued that “excessive regulation of the AI sector could kill a transformative industry just as it’s taking off,” and called for the EU to loosen safety and privacy restrictions that were shackling American companies and favoring entrenched interests.

Underneath these familiar themes of American dominance and freedom from government interference, however, lurked a more nuanced vision of tech regulation than many commentators recognized—but also some profound tensions.

Vance didn’t so much call for blanket deregulation as propose a distinction between bad regulation and good. To date, he complained, most AI regulation had been driven by a safety-ism that coddled consumers and a DEI agenda of information control that amounted to propaganda (as in AI-generated images of a black George Washington). But while calling for a rolling back of such rules in favor of “free speech,” Vance was equally adamant in calling for a regime of AI governance that fostered positive ends—specifically the interests of workers.

“We will always center American workers in our AI policy,” he declared. “We refuse to view AI as a purely disruptive technology that will inevitably automate away our labor force. We believe and we will fight for policies that ensure that AI is going to make our workers more productive.”

Such policies, of course, will require forms of regulation—closing off certain uses of AI that would simply destroy jobs and channeling innovation toward other uses that would leverage human skill and dignify human work. And they will, presumably, require aggressive and strategic government investment, rather than leaving innovation purely to the private sector. But this is an agenda of AI regulation that treats adults as grown-ups—as producers who need to be empowered with the best tools, not consumers who need to be sheltered from bad ideas.

Implicit in such an agenda, however, is the need to make careful discrimination between what is appropriate for adults and what is appropriate for children. AI regulations that would keep large language models from ever returning “racially insensitive” results are one thing. Regulations that would protect children from deepfake pornography are quite another. Vance observed as much in passing, but much more needs to be said. It’s not “safetyism” to highlight the real dangers of perhaps the most powerful technology in history.

Similarly, if we’re to use AI to enhance rather than replace human skill, we will need to ensure it is deployed in age-appropriate ways: A 30-year-old master programmer using ChatGPT to help write code is not the same as a 12-year-old using it to do his homework for him. Yet Vance spoke blandly of the need for schools to “teach students how to manage, how to supervise, and how to interact with AI-enabled tools.”

More fundamentally, a profound ambiguity lurks under the surface of Vance’s bid to preserve American AI dominance: Is it a national security imperative or a market imperative? On the one hand, the resemblances to the nuclear arms race are not hard to spot. Like nuclear energy, AI is one of the greatest technological breakthroughs in history, one with profound productive and profound destructive potential. As in the 1940s-'50s, the United States is at the cutting edge of the breakthrough, but an authoritarian geopolitical rival is close behind; once again, Europe is a battleground between the two. In both cases, therefore, too much hand-wringing about safety could have dire consequences: Better to put a dangerous product in American hands than let the industry be dominated by our enemies.

Unlike nuclear energy, however, AI can be implemented almost anywhere. Thus, while the U.S. government kept a pretty tight rein not just on nuclear weapons but even on private sector nuclear power, AI is being deployed across our economies in a breakneck competition for windfall profits. A true America First agenda for AI would require a subordination of market imperatives to national interests, but with Big Tech executives lining up outside the Oval Office, is that something the new administration is ready for?