Artificial intelligence is advancing faster than any technology in history. AI’s speed is not the only thing that makes it unique. Unlike with any past technology, AI executives are telling us that their goal is to replace human labor entirely.
They may not succeed. But the fact that they’re trying means that government needs to take it seriously and act now.
That’s why I’ve proposed the AI Dividend.
The enormous wealth AI companies are generating is built off our data, our ideas, our work. If AI transforms the economy and the role of human work in it, the gains should flow back to the public, not just to the people building the technology. The idea behind the AI Dividend is simple: if AI dramatically increases productivity and concentrates wealth, the American people share directly in those gains.
The AI Dividend is an insurance policy for the American workforce. It doesn’t assume the worst. It prepares for a range of outcomes. Just as we buy fire insurance not because we expect our house to burn down, but because we can’t afford the risk of being unprepared, the AI Dividend ensures that if AI dramatically reduces employment, Americans have a floor to stand on. The program is tied to clear economic triggers—sustained declines in labor force participation, wage compression in affected sectors, or rapid increases in AI-driven productivity without corresponding job growth—so that it activates based on real-world conditions, not political discretion. When those thresholds are met, revenue flows to three places: direct payments to Americans, investments in workforce transition and education, and public infrastructure to govern AI safely.
The funding mechanisms are designed so that as AI scales, so does the safety net. The first is a token tax: AI usage is measured in tokens, units of computation, and a modest tax on commercial AI consumption ties the funding directly to adoption. The second is equity participation in frontier AI firms. The federal government would take stakes in major AI companies structured as out-of-the-money warrants—the right to buy shares at a price that only pays off if these companies are wildly successful at replacing human labor, which is exactly when the dividend would be most needed. The third is reforming how we tax labor and capital. Right now our tax code is exactly backwards: companies can write off AI investments on their taxes, but pay additional taxes for hiring workers. If AI is substituting for labor rather than complementing it, then our tax code is actively subsidizing job elimination. We need to change that—reducing the ability to depreciate AI investment when it leads to less work, and making it cheaper to hire people than to replace them.
The window to act is closing. Once a small number of companies have accumulated extraordinary wealth and displaced workers across the economy, the political and practical window for creative policy closes. Demanding stakes in companies after they have already captured the value is far harder than building smart structures today, while the technology is still taking shape. This is not a bet that AI will necessarily cause mass unemployment. It is a recognition that the risks are high enough that preparation is essential—and that without proactive policy, the gains of the AI revolution will concentrate at the top while working people absorb the disruption.
I’ve already fought this fight in Albany, passing the RAISE Act, the nation’s strongest AI safety law, over the furious objections of AI executives determined to prevent regulation of their industry. In Congress, I’ll fight just as hard to make sure that all Americans have a stake—and security—in the AI future.