OpenAI published a 13-page policy paper on Monday calling for taxes on automated labor, a government-managed wealth fund seeded by AI companies, and pilot programs for a 32-hour workweek at full pay. The document, titled "Industrial Policy for the Intelligence Age," landed as Congress prepares to debate AI legislation and midterm campaigns gear up.
The company that just became the most valuable private tech firm in history is now asking the government to tax the thing it sells. Make of that what you will.
The money part
The core fiscal argument in the policy paper is straightforward: if AI displaces enough workers, payroll taxes collapse, and with them the funding for Social Security, Medicaid, and SNAP. OpenAI's proposed fix is shifting the tax base toward capital gains and corporate income, plus some form of levy on automated labor.
The company does not propose a specific tax rate. That's convenient, given that the Trump administration cut corporate taxes from 35% to 21% during the first term, and OpenAI president Greg Brockman has personally donated millions to the president. As TechCrunch noted, this kind of capital gains proposal is exactly what pushed Marc Andreessen toward Trump after Biden floated taxing unrealized gains in 2024.
Then there's the Public Wealth Fund, the paper's most ambitious proposal. Every American would get a direct ownership stake in AI-driven economic growth through a nationally managed fund, partly capitalized by AI companies themselves. Returns distributed to citizens regardless of their existing wealth. It reads like a sovereign wealth fund with extra steps, and the document says nothing about how it would actually be financed beyond suggesting policymakers and AI companies figure that out together.
32 hours and you're done
OpenAI wants employers and unions to pilot a four-day workweek at full pay, framing the shorter hours as an "efficiency dividend." If AI cuts your company's operating costs, some of that surplus should become your time, not just shareholder returns. The paper also suggests companies increase retirement contributions, expand healthcare coverage, and subsidize childcare.
All corporate responsibilities, though. Not government ones. TechCrunch's Rebecca Bellan caught the gap: if automation eliminates your job entirely, your employer-subsidized benefits vanish with it. OpenAI separately proposes portable benefit accounts that follow workers across jobs, but those still depend on employer contributions and fall well short of universal coverage.
So who's this for?
The blog post accompanying the paper frames everything as "intentionally early and exploratory," which is corporate for "we know none of this will happen as written." OpenAI is offering up to $100,000 in grants and $1 million in API credits for policy research building on these ideas, and hosting a workshop in Washington in May.
The timing matters more than the specifics. This drops six months after Anthropic's blueprint on AI-driven disruption and right as the administration shapes a national AI framework. Sam Altman told Axios that a major AI-enabled cyberattack is "totally possible" within the next year and that AI-created pathogens are "no longer theoretical." The paper itself includes containment playbooks for AI systems that can self-replicate and operate autonomously, systems that, in its own words, "cannot be easily recalled."
An $852 billion company publicly calling for taxes on its own technology is, at minimum, a signal that the people building this stuff expect disruption soon, not in some hazy future. Whether these proposals represent genuine policy advocacy or positioning for regulatory negotiations is the question that will follow this document around Washington all spring.
OpenAI is collecting feedback at [email protected]. The May workshop has not yet been given a specific date.
