OpenAI published a 13-page policy blueprint on Monday proposing that governments tax automated labor, create a national wealth fund giving every citizen a direct stake in AI profits, and run government-sponsored 32-hour workweek pilots. The document - titled "Industrial Policy for the Intelligence Age: Ideas to Keep People First" - frames the proposals as analogous to the New Deal: a political and economic response to a technology transition that's already displacing workers before any safety net exists to catch them.

The timing is not accidental. OpenAI closed a $110 billion funding round at an $850 billion valuation earlier this year and is moving toward an IPO. The company that's arguably most responsible for the current wave of AI-driven job disruption is now formally lobbying for redistributive economic policy - and doing so publicly, eight weeks before its Washington workshop opens in May.

What OpenAI Is Actually Proposing

The document covers 20 distinct policy ideas across two sections: building an open economy and building a resilient society. The proposals worth watching fall into five clusters.

The robot tax

The document doesn't specify a rate. The language is careful: it says the tax base may need to shift from payroll toward capital gains, corporate income, and "targeted measures on sustained AI-driven returns" - with "taxes related to automated labor" listed as one possible tool. Sam Altman, in an Axios interview published the same day, called large tax changes "in the Overton window, but near the edges."

That hedging matters. OpenAI is not calling for a specific robot tax. It's putting the concept on the table and gesturing at the problem it would solve: as payroll shrinks and profits concentrate in AI companies, Social Security, Medicaid, SNAP, and housing assistance lose their funding base. The document names this directly.

The wealth fund

The proposal models a national wealth fund on Alaska's Permanent Fund, which distributes annual oil revenue dividends to state residents. OpenAI proposes that AI companies and policymakers jointly seed a federal version and invest it in "diversified, long-term assets that capture growth in both AI companies and the broader set of firms adopting AI." Returns go directly to citizens. No seeding amount is specified.

The fund would function as a forced equity stake for Americans who aren't currently invested in financial markets - which is most of them. It doesn't require new legislation to be technically possible, but it requires a level of coordination between Congress and private companies that has no precedent in US tech policy.

Sam Altman at TED2025 in April 2025, where he discussed AI's economic impact and the need for new policy frameworks. Source: blog.ted.com

The workweek pilot

The four-day workweek proposal is the most concrete in the document. OpenAI calls for encouraging employers and unions to run "time-bound 32-hour/four-day workweek pilots with no loss in pay that hold output and service levels constant," then convert the reclaimed hours into a permanent shorter week or bankable paid time off. The document also proposes increased retirement matching and employer-covered childcare as part of the same efficiency-dividend package.

Automatic safety nets

This section proposes building real-time measurement of AI's impact on work and wages, then defining a package of temporary expanded benefits - unemployment insurance, cash assistance, wage insurance, training vouchers - that activates automatically when displacement metrics hit preset thresholds. When conditions improve, the package phases out. The design explicitly avoids permanent program expansion.

The Counter-Argument

Critics will note that OpenAI is the entity that most benefits from the policies it's lobbying against. The document itself acknowledges this: "There is also a risk that the economic gains concentrate within a small number of firms like OpenAI." That admission, buried on page seven, is more candid than most corporate policy papers get. It doesn't, however, resolve the tension.

The proposals require government implementation. OpenAI isn't in a position to unilaterally impose a robot tax, fund a wealth fund, or legislate a shorter workweek. What the company can do - and is doing - is shift the political conversation in a direction that positions it as a responsible actor ahead of its IPO and ahead of what it describes as a coming transition to superintelligence.

"Some will be good. Some will be bad. But we do feel a sense of urgency. And we want to see the debate of these issues really start to happen with seriousness."

Sam Altman, speaking to Axios, April 6, 2026

The same interview had Altman calling an AI-enabled major cyberattack within the next year "totally possible" and describing AI-enabled novel pathogens as "no longer theoretical." The policy document doesn't engage with those risks at the same level of specificity. It treats economic disruption as the primary near-term policy challenge - which is at least partially a framing choice that benefits a company selling productivity tools.

There is also a structural critique. Researcher Toby Walsh has argued that robot taxes, while intuitive, are difficult to define - automation has always displaced some jobs while creating others, and drawing a line between AI-assisted work and AI-replaced work is legally and economically contested. OpenAI's document acknowledges this implicitly by keeping the language vague.

Manufacturing is among the sectors OpenAI identifies as most exposed to AI displacement, though the document focuses more on administrative and cognitive work. Source: unsplash.com

The Displacement Data It Cites

OpenAI's framing aligns with independent research. Andrej Karpathy's job exposure analysis found that 42% of US occupations score 7 or higher on a 0-10 AI exposure scale, covering 59.9 million workers and $3.7 trillion in wages. Challenger Gray data cited by earlier reporting showed AI-attributed layoffs hit 55,000 in 2025, a 12x increase from two years prior. Anthropic's own job risk study found young workers and recent college graduates feel the exposure most acutely.

These numbers are real. The debate isn't over whether AI is displacing workers but how fast, how concentrated, and whether any safety net can respond at the right speed. OpenAI's adaptive safety net proposal - automatic triggers rather than permanent expansion - is one of the more technically sound ideas in the document because it matches the policy response to the speed of the disruption rather than locking in a permanent program.

What the Market Is Missing

The market reads documents like this one as PR. That may be partly right - OpenAI is heading into an IPO and wants to look like a responsible actor. But the policy proposals themselves deserve evaluation on the merits, and several of them are more serious than they appear in a headline.

The adaptive safety net mechanism - automatic activation at displacement thresholds, automatic phase-out when conditions improve - solves a real problem in social policy: the inability of legislated programs to respond quickly to economic shocks. The portable benefits proposal addresses a genuine structural failure in US labor markets where benefits tied to single employers lock workers out of entrepreneurship and mobility.

The wealth fund and robot tax are the weakest proposals, not because the problem they address is wrong but because the mechanics are undefined. Until OpenAI specifies what counts as "automated labor" for tax purposes and how a nationally seeded wealth fund gets capitalized without congressional action, those proposals are political positioning, not policy.

OpenAI is right that the transition is already happening. It is also in the unusual position of simultaneously accelerating that transition and proposing to cushion it. The 13 pages are worth reading in full - the PDF is publicly accessible at OpenAI's policy site - not because they represent a final plan but because they're the most detailed thing any frontier AI lab has published about what it thinks governments should actually do.

Sources:

• OpenAI - Industrial Policy for the Intelligence Age (PDF)
• OpenAI - Global Affairs page
• TechCrunch - OpenAI's vision for the AI economy
• Newsweek - Sam Altman proposes robot tax
• The Next Web - OpenAI robot taxes wealth fund