OpenAI calls for robot taxes and a public wealth fund

OpenAI urges robot taxes and a public wealth fund, company floats four-day week pilots as AI job fears spread, the firm selling automation also asks government to socialise the adjustment bill

OpenAI has published a new set of policy ideas it says could soften the economic shock of advanced AI, including a “robot tax,” a public wealth fund, and experiments with shorter working weeks. The proposals, reported by Business Insider, arrive as companies deploy AI tools more aggressively and as lawmakers face pressure to show they have a plan for job displacement.

The package is framed as preparation for “AI-fuelled economic upheaval,” with OpenAI pointing to the possibility that automation will remove or compress large categories of white-collar work. The measures it highlights are familiar in the tech-policy circuit—taxing capital that substitutes for labour, recycling proceeds into broad-based payouts, and using working-time reductions to spread remaining work. Sam Altman has previously backed ideas such as universal basic income, and OpenAI has run small-scale UBI-related research through affiliated initiatives; the new document positions similar concepts as mainstream contingency planning rather than speculative philanthropy.

The politics of a robot tax start with a definitional problem: what counts as a “robot” when the productivity gain comes from software sitting behind an API. Tax systems handle machines tolerably well when they are bought as capital equipment and depreciated; they handle subscription software and model access far less cleanly. A levy that targets “AI” risks becoming either a compliance industry for large incumbents—who can document and classify usage—or a blunt surcharge on any firm modernising its processes. The more precise the tax, the easier it is to avoid by relabelling tools; the broader it is, the more it begins to look like a general tax on investment.

A public wealth fund has a different attraction: it turns the question into ownership rather than measurement. If automation concentrates profits in a handful of firms with scarce compute, data, and distribution, a fund can buy stakes and distribute returns. But that requires the state to accumulate capital at scale, decide what to buy, and resist the temptation to use the fund as an off-balance-sheet industrial policy tool. In practice, such vehicles often end up holding the same large listed companies the private sector already owns—while the political fight shifts to who controls the fund and what counts as a legitimate payout.

The four-day week proposal is the least fiscal and the most operational: it asks employers to reorganise work rather than pay a new tax. Yet it too depends on who absorbs the cost. In sectors where output is tied to hours—healthcare, logistics, retail—shorter weeks typically mean more staff or more overtime, not free efficiency. In sectors where output is tied to projects and deadlines, reduced hours can become a paper policy that quietly shifts work into evenings and contractors. The more AI raises measured productivity, the stronger the case that firms can keep output while cutting hours; the more AI raises competitive pressure, the stronger the case that firms will use the gains to undercut rivals instead.

OpenAI’s list lands at an awkward moment for any company proposing new taxes: it is simultaneously a supplier of the tools that replace labour and a beneficiary of the profits that follow. The policy debate it is trying to shape is, in effect, about how to allocate the surplus from systems it builds.

Business Insider notes that OpenAI’s recommendations include both a robot tax and a public wealth fund—two different ways of turning private automation gains into public revenue.