Bernie Sanders isn’t afraid to ask the big questions. Writing in the New York Times, he asked: “Will the future of humanity be determined by a handful of billionaires who have promoted and developed AI, with virtually no democratic input?”
It’s a fair question. The concentration of power, wealth, and control among tech oligarchs is probably the most urgent risk AI poses to democracy. Sanders and I agree on that much.
Where we disagree is on the solution. Sanders wants to create a US sovereign wealth fund by taking 50% stock in AI companies like Anthropic, OpenAI, and xAI. The idea is twofold: establish democratic control through voting shares, and return the economic rewards of soaring AI valuations to the public.
Those are worthy goals. But public ownership of these companies would entangle corporate profit with the public interest in dangerous ways. If growing Nvidia from $5 trillion to $10 trillion also doubles the value of the sovereign wealth fund’s stake, you can bet fund managers will support chip sales with the same zeal as private investors.
We’ve seen this before. Norway’s sovereign wealth fund owns massive stakes in oil companies, and it hasn’t steered them toward pro-environmental policies. If anything, the government’s dependence on those companies has inhibited climate action. Public employee pension funds face the same criticism — fiduciary duty to generate wealth overwhelms any intention to direct corporate behavior.
There’s a better approach: separate the two goals. Use taxation to share private rewards with society — Elizabeth Warren’s proposed excise tax on datacenter energy use, or Mark Cuban’s AI token tax. And for reshaping AI in the public interest, create a public AI option. Governments should develop and operate AI models under democratic control, providing a competitive baseline that private offerings must meet.
Switzerland is already doing this. Apertus is a large language model built by Swiss public servants and university researchers, using licensed training data and public supercomputing infrastructure. It doesn’t beat OpenAI on benchmarks, but it blows them out of the water on transparency and regulatory compliance.
Sanders is a savvy political operator, so why pursue the sovereign wealth fund strategy? Probably because the AI billionaires themselves are open to it — they believe that for every dollar ceded to government stock expropriation, they’ll get back more in favorable policy. That should tell us something.
Energy taxation and public AI represent a genuine chance to influence corporate behavior toward the public interest. We should push for those instead.
