June 3 2026 - 6:03pm
There’s one op-ed in the New York Times this week that’s unlike the others. Written by Vermont senator Bernie Sanders, it argues that AI should be treated as a public resource, with its gains redistributed. Sanders proposes a one-off 50% tax on the equity of major AI firms, with the proceeds used to create a sovereign wealth fund. The model, he notes, is similar to those in Norway and Alaska, where oil revenues are pooled for public benefit.
The op-ed also serves as a prelude to forthcoming legislation — the American AI Sovereign Wealth Fund Act — which Sanders says will soon be introduced in Congress.
To Sanders’s supporters, the bill looks like a win-win. The AI sector already generates roughly $400 billion in revenue and is projected to scale into the trillions. Much of that would accrue as profit to a small group of tech firms — wealth they could, in this view, easily forgo. Rather than entrenching inequality or what some economists warn could become a “permanent underclass”, those gains would instead be redistributed more broadly across society.
Unfortunately, however, this bill is never going to pass. If the US were a socialist country, it would be no issue, but the US is not a socialist country. The government isn’t in the habit of nationalizing major industries, and there is no clear dividing line here for how the US would be justified in nationalizing AI as opposed to nationalizing any other industry. Sanders’s argument — that “AI is built on our collective intelligence” — is threadbare. This justification might carry some intuitive force, but it is far more persuasive in the case of oil, a natural resource that, in a basic sense, belongs to the territory from which it is extracted.
What’s more, it is not as if AI was trained only on American citizens, or even equally on the data of all American citizens such that they would be entitled to an equal dividend from it. And it remains a complex legal question, currently wending its way through the courts, whether AI’s use of news articles in its training data constitutes copyright infringement or the “transformative” development of a categorically different product.
But perhaps it speaks to a deeper truth of the AI era: hype often matters more than substance. The mere fact of Sanders introducing the bill and publishing an op-ed in the New York Times signals a shift in the relationship between technology and politics, regardless of whether the bill ever reaches a Senate vote. Big Tech is inevitably going to become a major campaign issue. In fact, it is surprising it hasn’t already. The average American spends five hours a day on their smartphone and nearly 10 on a screen: this is where people actually live, and it shapes daily life more than culture-war issues such as transgender athletes or unisex bathrooms.
There was arguably an opening to turn this into a wedge issue through antitrust campaigns against Amazon or Facebook, as Lina Khan hinted at during her tenure at the FTC. Yet no one has quite managed to convert Big Tech anxiety into a durable electoral message. Sanders, who has strong political instincts, may have found a way in.
To campaign against Big Tech is to find a new, empty space in the political map, and a potential horseshoe. MAGA nurses a deep resentment of Big Tech and its liberal Silicon Valley culture. Meanwhile, as Sanders has intuited, there’s a potent populist-progressive line of attack against AI waiting to be exploited. Whether his proposal is actually feasible may be beside the point — it’s a platform, not a policy.
