Under one plausible theory of technological development, artificial intelligence will turn the economy into a dystopian split screen. As AI becomes capable of replacing an ever larger share of human work, the companies selling it will grow astronomically, turning a handful of investors and technologists into the richest individuals in human history while creating a permanent underclass of former workers replaced by machines.
To head off that scenario and the consequences it could have for our political and social order, a growing number of figures in both Washington and Silicon Valley have begun advocating a policy idea known as “universal basic capital.” The specific proposals vary, but the basic intuition is simple: If the wealth generated by artificial intelligence will accrue only to the owners of AI companies, then the way to avoid catastrophic levels of inequality is to give everyone an ownership stake in those companies. The idea has been endorsed by figures as ideologically diverse as Bernie Sanders, Gavin Newsom, Steve Bannon, Sam Altman, and even Donald Trump.
Designed the right way, the policy could provide a much-needed hedge against a future AI dystopia. It wouldn’t prevent workers from losing their jobs to machines, but it could guarantee that they share in at least some of the economic upside of the technology. Unfortunately, the version of the idea with the most momentum is very much not designed the right way. Were it to be implemented, it would invite its own kind of disaster.
As the name suggests, universal basic capital is a new twist on a slightly older concept: universal basic income. Under a UBI scheme, the government would raise taxes on the profits of wealthy companies and then distribute the proceeds in the form of checks mailed out to every American. During the 2020 Democratic primary, Andrew Yang became a political celebrity by running a single-issue campaign devoted to implementing the policy before mass AI-triggered job loss.
But massively raising both taxes and welfare payments is a political nonstarter. Universal basic capital avoids that problem, at least in theory. Instead of a monthly check, every American would be given an account containing a chunk of equity invested in a broad portfolio. The value of that account would compound over time as the market rises. Crucially, whether or not people receive that money wouldn’t depend on future tax revenues or be vulnerable to shifting political winds; once you have been given the equity, it’s yours. (The accounts could be funded in several ways, such as a onetime levy on the market capitalization of large companies, to be paid for in stock instead of cash.)
[Rogé Karma: So, about that AI bubble]
Most important, the policy might actually have a puncher’s chance at gaining bipartisan support. The One Big Beautiful Bill, which passed last year with near-unanimous Republican support in Congress, already established a micro-version of universal basic capital in the form of “Trump Accounts” that provide every child born from 2025 to 2028 with a $1,000 brokerage account at birth. By way of explaining the policy’s conservative appeal, Republican Senator Ted Cruz argued that universal stock ownership would “make every new child a capitalist.” In a December op-ed, the Republican Ohio gubernatorial candidate Vivek Ramaswamy proposedbuilding on that idea for the AI age by establishing an “American dream birthright” in the form of a $10,000 account invested in the S&P 500. “Instead of lambasting millionaires,” Ramaswamy wrote of young Americans, “they would be on the way to becoming millionaires.”
“The great strength of universal basic capital is that it is so hard to pin down ideologically,” Nathan Gardels, a co-founder of the Berggruen Institute, a tech-focused think tank, told me. “It doesn’t fall neatly into the kinds of partisan divides we’re used to dealing with on these issues.”
The other case for universal basic capital is that it would have benefits even if the worst fears of an AI job apocalypse never come to pass. AI might very well create a lot of wealth without putting huge swaths of the population out of work. Even in that future, however, broad capital ownership would go a long way toward making sure that everyone shares in the benefits of technological progress. The richest 10 percent of Americans currently own about 90 percent of stocks, while the bottom half owns less than 1 percent. If that distribution were to remain constant, then the AI boom would produce a socially and politically toxic level of wealth inequality even if it didn’t destroy a single job.
And even if AI turns out to be a financial bubble that results in a stock-market correction, ordinary people would still be better off holding some assets, which will eventually rise in value, than none. “This is what I would call ‘no-regrets policy,’” David Autor, an MIT economist, told me. “It’s a good idea in just about any scenario you can think of.”
Unfortunately, the proposal that is getting the most attention right now is much more complicated, and much worse, than just giving everyone an AI stock-market account.
In a recent New York Times op-ed, Bernie Sanders proposed requiring AI companies, such as OpenAI, Anthropic, and xAI, to hand over a 50 percent equity stake in their business to the federal government. Those shares would be placed into a public “sovereign wealth fund” that would grow in value over time and be used to make direct payments to the American public and to fund public programs. As evidence for how this would work, Sanders points to Norway, which began placing the returns from its oil wealth into a sovereign wealth fund in 1996 and has since watched that fund grow to $2.2 trillion, about four times as large as the entire Norwegian economy. “This legislation would guarantee the trillions of dollars potentially generated by A.I. are used to improve the lives of all of us—not simply to make the richest people in the world even richer,” Sanders wrote. He has since introduced the proposal in the form of a bill called the American A.I. Sovereign Wealth Fund Act.
Unlike the simpler AI-accounts model, the Sanders approach would grant the federal government voting rights and representation on each AI company’s board, giving it a direct say over those companies’ day-to-day operations and key decisions. To proponents, that’s a big advantage. “This is about giving the public a seat at the table,” Sarah Polcz, a law professor at UC Davis whose AI-wealth-fund proposal influenced the Sanders bill, told me. “If you just give everyone individual accounts, you miss a big opportunity to shape how the technology develops and who it serves.”
You might expect AI companies (which would prefer not to have their assets seized) and Republicans (who traditionally hate government ownership) to protest. Instead, the opposite has happened. Shortly after Sanders unveiled his proposal, OpenAI CEO Sam Altman, whose company had recently proposed the creation of a “Public Wealth Fund,” requested a meeting with the senator. In Altman’s telling, the two agreed on the broad thrust of the policy even as they disagreed over specific details, such as the 50 percent number. Days later, Trump told reporters that his administration was working on a potential partnership with AI companies whereby “pieces” of the companies “could be given to the American public.” When reporters noted the similarities between the president’s suggestion and the Sanders proposal, Trump replied that “as far as economics is concerned,” he and Sanders “aren’t that far apart.” The president later said that he would be inviting the heads of the AI companies to the White House to discuss the idea.
Perhaps this is all just talk—a way for Altman and Trump to earn some goodwill with the American people by claiming to support an idea that will never happen. But an even more worrying possibility is that the AI companies and Trump are serious about the idea because they expect to gain from it. If the U.S. government owned huge portions of, say, OpenAI and Anthropic, it would have a strong incentive to do everything in its power to make sure those firms were financially successful. That might mean gutting labor or safety standards that would delay the technology’s rollout, ignoring anticompetitive acts, or providing favors in the form of cheap loans or lucrative government contracts. If the AI sector turned out to be a bubble, as many fear, the companies could likely count on a government bailout. “The federal government is really the only entity powerful enough to be a real check on these companies,” Samuel Hammond, the chief economist at the Foundation for American Innovation, a center-right think tank, told me. “If they become joined at the hip, that check goes away. It can easily become a form of regulatory capture.”
[Vivian Salama: J. D. Vance’s AI doctrine]
These problems could be avoided with carefully constructed restrictions to mitigate conflicts of interest. Norway, for instance, has been a global leader in the transition away from fossil fuels and toward green energy despite the fact that its fund is seeded by oil revenues. But America isn’t Norway. Although Sanders calls for the wealth fund to be managed by an “independent commission,” the idea of a massive new state-run enterprise maintaining operational independence from political actors in the Trump era is laughable. (Just this week, the Supreme Court ruled that the president could remove the leaders of “independent agencies” at will.) A sovereign wealth fund would functionally hand Trump a giant pot of money that he could use to enrich himself and his family, hand out favors to political allies, and force business leaders to bend the knee. It would also give him control over how the technology is developed and deployed. This would radically alter the balance of power between Washington and Silicon Valley. Earlier this year, Anthropic refused to allow the U.S. military to use its technology for domestic surveillance or autonomous weapons. That kind of decision would no longer be possible if the government owned a controlling stake in the company. “We’ve already seen what happens when Trump is able to tell TV stations who they should have as a late-night host or how to run their news shows,” Dean Baker, a senior economist at the Center for Economic and Policy Research, told me. “Do you really want to see what happens when we give him the power to run large chunks of corporate America?”
Such an outcome would be tragic in the classic sense. Thanks to a twist of political fate, there is some semblance of bipartisan agreement around how to address an existential problem before it even happens. That kind of opportunity might not come along again for a long time, if ever. It would be a shame to waste it.