Here is a sentence that would have been incomprehensible two years ago: a New York state assemblymember running for Congress just published a detailed policy memo proposing that the federal government tax AI tokens and use the revenue to send checks directly to Americans whose jobs get automated away.

Alex Bores calls it the "AI Dividend." Axios broke the story. Then Ezra Klein put him on the most influential opinion podcast in the country to pitch it. Then Gizmodo, Yahoo Finance, and Breitbart all picked it up the same week. The proposal is now the most specific, most widely discussed AI-and-jobs policy in the 2026 midterm cycle.

And the AI industry is terrified of it.

What the AI Dividend Actually Is

The proposal is smarter than it sounds. It is not a blanket tax on AI companies. It is a trigger-based system. The AI Dividend only activates when the government determines that AI has begun to "meaningfully" displace workers. Bores defines meaningful as one of three conditions: a persistent decline in labor force participation, wage compression in AI-exposed sectors, or a notable increase in AI-driven productivity without parallel job growth.

Once triggered, the money comes from what Bores calls "AI-linked revenue mechanisms." The two headliners: a modest tax on AI token consumption, tied to the volume of compute used, and the option for the federal government to take equity stakes in major AI firms. The revenue funds three things: direct payments to displaced workers, workforce retraining, and government AI oversight capacity.

The key line from his policy memo: "If AI can substitute for labor rather than complement it, then our tax code is actively subsidizing job elimination. We encourage companies to invest in AI by making it cheaper through tax breaks, while taxing the wages of the workers being displaced. Taxing the thing that is growing, AI, rather than the thing that is shrinking, wages, is simply sound fiscal management."

That is a clean argument. It is also the kind of argument that makes Sam Altman's lobbyists reach for the phone.

Why OpenAI and Palantir Want Him Gone

Bores is not just any candidate with an AI opinion. He co-sponsored the RAISE Act, New York's landmark state-level AI safety legislation. He has been the most vocal AI hawk in any state legislature for over a year. That track record made him the prime target of "Leading The Future," a super PAC bankrolled by Andreessen Horowitz, OpenAI president Greg Brockman, Palantir co-founder Joe Lonsdale, and the AI search engine Perplexity.

Read that list again. The president of OpenAI is personally funding opposition to a candidate whose proposal would tax OpenAI's product. Palantir's co-founder is doing the same. These are not abstract policy disagreements. These are the companies that would pay the tax if the AI Dividend becomes law.

The Ezra Klein podcast amplified the story further. The episode, titled "Why Are Palantir and OpenAI Scared of Alex Bores?", laid out the case that AI companies are spending aggressively on midterm races specifically to prevent candidates like Bores from reaching Congress. Not because they think his proposals are bad policy. Because they think his proposals might actually pass.

The Context That Makes This Land

This proposal does not exist in a vacuum. It arrives during the same week that tech layoffs in 2026 crossed the 100,000 mark. Meta confirmed 8,000 cuts on May 20. Microsoft offered voluntary buyouts for the first time in 51 years. Anthropic's own research survey of 81,000 Claude users found that the people who use AI the most are the most afraid of losing their jobs to it.

The Bores memo addresses this directly: "For decades, new technologies have created more jobs than they destroyed. But AI is different. For the first time, the people building the technology are explicitly trying to automate all human labor. They may not succeed. But the fact that they are trying means government needs to take the possibility seriously."

He is not wrong. Dario Amodei told his own staff that AI could automate 50% of jobs within two years. Sam Altman has floated Universal Basic Income, Universal Basic Compute, and a four-day work week as potential responses. Elon Musk pinned a "Universal High Income" proposal to his X profile. The people building these systems are telling you, explicitly, that mass displacement is coming. Bores is the first candidate to take them at their word and propose a mechanism to pay for it.

My Take

The AI Dividend is the first AI-and-jobs proposal I have seen from a politician that is not embarrassingly vague. It has a trigger mechanism, a funding source, and a delivery pipeline. You can argue about the thresholds. You can argue about whether token taxation is the right mechanism. You can argue about whether the government should take equity stakes in AI companies. Those are real policy debates worth having.

What you cannot argue is that AI companies are spending millions to defeat a candidate in a House race. That is the tell. If the proposal were toothless, they would ignore it. They are not ignoring it.

The 2026 midterms are shaping up to be the first American election where AI policy is a real campaign issue. Not a talking point on a debate stage. Not a throwaway line in a stump speech. A specific proposal with specific funding mechanisms that specific companies are spending specific money to kill.

Whether Bores wins his race or not, the AI Dividend just became the benchmark. Every other candidate with an AI platform is now going to be measured against it. And every AI company that funded opposition to it just told you exactly how seriously they take the possibility that governments might actually make them share the wealth.