What happened (and what OpenAI is actually trying to do)

OpenAI published a policy paper, “Industrial Policy for the Intelligence Age,” pitching a bundle of ideas for what happens when “superintelligent” systems start rearranging work, taxes, and who gets paid. The list includes shifting tax burden from labour to capital, a possible robot tax, a public wealth fund that pays citizens a stake in AI returns, and support for a four-day workweek without pay cuts.

On paper, it’s “people-first.” In practice, it’s also agenda-setting. If you’re the company racing to build the thing that scrambles labour markets, you don’t want to be regulated by a committee that learned what an API is last Tuesday.

The non-obvious part: this is a fight over who owns the narrative of “inevitable”

There are two competing stories about AI disruption. Story A: “progress” happens, politics reacts, and everyone muddles through. Story B: disruption happens, and the winners write the compensation mechanisms before the losers even get a ticket number.

OpenAI is trying to make Story B sound like responsible leadership. The proposals are broad enough to look benevolent, and vague enough to keep optionality. That’s not a conspiracy. It’s the normal behaviour of any powerful actor trying to shape a market structure story before it becomes a liability story.

Where the paper is useful

• It admits the tax base problem. If profits and capital gains rise while payroll taxes shrink, the programs funded by wage income get hollowed out. That is the kind of boring mechanical detail that actually matters.
• It ties the AI economy to energy and infrastructure. The paper doesn’t just do redistribution, it also pushes for faster power buildouts and subsidies. Translation: the “AI future” is a grid and permitting story, not a chat UI story.
• It treats access as a strategic risk. OpenAI argues AI should be “utility-like,” which is another way of saying: don’t let regulation create a world where only incumbents can comply, except also please don’t regulate us into a utility.

Where it’s still vibes (and why critics aren’t wrong)

Even friendly readers point out the same gap: naming the ideas is not the same as building mechanisms. “Public wealth fund” is not a bill. “Portable benefits” is not an implementation plan. “Four-day week subsidy” is not a political coalition.

And there’s an unavoidable trust problem. The labs most invested in the outcome are also the least neutral parties. That does not mean you ignore them. It means you treat their policy papers like you treat their benchmarks: as marketing until proven otherwise.