OpenAI Calls for Robot Taxes, a National Wealth Fund, and a Four-Day Workweek

OpenAI has never been shy about its ambitions. But a 13-page policy document published Monday marks a new kind of ambition: not just to build superintelligence, but to tell governments how to govern the world it will create. The document, titled “Economic Policy for the Intelligence Age,” lays out a detailed framework for how OpenAI believes societies should manage the economic disruption that frontier AI will inevitably cause — and the proposals are far more specific, and far more politically charged, than anything the company has previously published.

The timing is not accidental. OpenAI has reportedly begun preliminary discussions with investment banks ahead of a potential IPO as early as late 2026. Publishing a policy vision that positions the company as a thoughtful steward of the AI economy — rather than a corporation racing for dominance — is also a reputational and regulatory strategy. Whether that framing is genuine or calculated, it is now the public record OpenAI will be measured against.

The Three Pillars

The document centers on three stated goals: distributing AI-driven prosperity broadly, building safeguards against systemic risks, and ensuring that access to AI capabilities doesn’t become so concentrated that economic power and opportunity calcify around a handful of companies and countries.

To achieve these goals, OpenAI proposes three headline policies.

A Robot Tax. OpenAI endorses shifting the tax burden from human labor to capital — specifically, automated systems that replace workers. This is not a new idea. Bill Gates proposed something similar in 2017. But OpenAI’s endorsement carries different weight: this is the company whose products are arguably the most proximate cause of white-collar automation at scale. The proposal includes taxes on revenue generated through AI-assisted labor, and mechanisms for those revenues to fund retraining programs and social safety nets.

A Public Wealth Fund. At the document’s center is a proposal for governments to create large, state-backed investment funds that hold equity stakes in AI companies and infrastructure. The analogy OpenAI draws is to sovereign wealth funds like Norway’s Government Pension Fund Global — accumulated national assets that generate returns for broad populations. The AI version would be seeded partly through equity stakes taken when AI companies receive government support, regulatory approval, or access to public data. Citizens would receive dividends as AI productivity grows.

A Four-Day Workweek. OpenAI’s most headline-friendly proposal is a transition to a 32-hour workweek with no reduction in pay. The logic is simple: if AI-driven productivity gains make it possible to produce the same or greater economic output with less human labor, the gains should be distributed partly as time rather than only as corporate profit. OpenAI draws on historical precedent — the shift from six to five working days in the early 20th century — to argue that workweek compression is a natural response to technological productivity gains.

Automatic Safety Nets

Beyond the three headline proposals, the document outlines a framework for automatic stabilizers tied to measured AI-related job displacement. Rather than waiting for Congress or Parliament to respond to layoffs, the proposal would pre-commit governments to deploying income support, wage insurance, and direct cash payments once displacement metrics cross defined thresholds. The model is explicitly borrowed from the U.S. system of automatic unemployment insurance expansion during recessions — applied to a permanent structural shift rather than a cyclical downturn.

This is arguably the most technically interesting part of the proposal, because it acknowledges something that most corporate AI commentary avoids: the displacement may not be temporary. Standard economic theory holds that technological unemployment is transitional — workers retrain and find new roles in new industries. OpenAI’s framing allows for the possibility that the transition period is long, or that some categories of work don’t come back at all, and proposes infrastructure to manage that outcome rather than simply assuming it away.

The Self-Interest Problem

The document has already attracted pointed criticism, most of it focused on a single uncomfortable fact: every major policy OpenAI is proposing would, in different ways, benefit OpenAI.

A robot tax that funds retraining programs reduces political backlash against AI adoption. A public wealth fund that holds equity in AI companies effectively makes the public a stakeholder in OpenAI’s success — and reduces the incentive for governments to regulate it harshly. A four-day workweek driven by AI productivity validates the claim that AI tools are producing real economic gains, a narrative OpenAI has strong commercial interest in promoting.

Critics have also noted that the wealth fund proposal is agnostic about corporate structure. OpenAI has been in a multi-year legal and structural restructuring away from its original nonprofit status; a government fund that takes equity stakes in AI companies would treat OpenAI’s for-profit entity the same as any other tech corporation, effectively ratifying the restructuring by making it the relevant unit of analysis for public investment.

None of this necessarily means the policies are wrong. Robot taxes and public wealth funds have been debated seriously by economists across the political spectrum for years. The four-day workweek has already been piloted with promising results in several European countries. But the source of these proposals — a company that is, by its own admission, the leading developer of the technology that makes them necessary — creates a legitimacy problem that the document does not fully address.

What Comes Next

OpenAI has shared the document with the White House economic policy team, the European Commission’s DG CONNECT, and several national AI advisory bodies. It is framed explicitly as a conversation-starter, not a legislative draft. Sam Altman has reportedly briefed select members of Congress on its contents.

The proposals are unlikely to become law quickly anywhere. Tax policy is contentious, workweek legislation faces significant employer resistance in most economies, and public wealth fund design is complex in ways that a 13-page document cannot resolve. But the document has already shifted the Overton window in one important direction: it is now on record that the leading AI company in the world believes AI-driven displacement is real enough to require structural policy responses — not just “learn to code” platitudes.

For an industry that has spent years insisting that AI will mostly create new jobs rather than eliminate existing ones, that admission is more consequential than any specific policy proposal in the document.