Last week, OpenAI released a set of AI economic policy recommendations containing three core proposals: establishing a public wealth fund, taxing AI companies, and promoting a four-day work week.
The move itself is somewhat intriguing: a company valued at $852 billion, with an annualized revenue exceeding $25 billion, and preparing for an IPO, is proactively advocating for its own tax increase and distributing profits to the general public.
Three policy proposals, each with a distinct focus
Public Wealth Fund: Automatically grant all U.S. citizens shares in AI companies, allowing them to receive dividends from the growth of the AI economy. The logic is similar to a sovereign wealth fund, but directed at individuals. Alaska's oil dividend plan and Norway's sovereign wealth fund follow similar paths.
Robot Tax (AI Profit Tax): Tax profits generated by AI, with the revenue used to fund social security programs and compensate for potential job losses. Specific tax rates and collection methods were not detailed.
Four-Day Work Week Subsidies: Use fiscal subsidies to encourage companies to shorten standard working hours, thereby addressing labor market pressures brought by AI. France, Germany, and Iceland have conducted similar experiments, with mixed results.
Sam Altman's recent policy moves
This document is not an isolated effort. Altman has recently appeared frequently in policy settings, including speaking at Federal Reserve events about AI's impact on the financial system. OpenAI also simultaneously launched an AI safety researcher fellowship program, opening up safety and alignment research.
The timing is also delicate: OpenAI just completed a $122 billion funding round, reaching a valuation of $852 billion, with annualized revenue exceeding $25 billion, and is preparing for an IPO later this year. At this scale, proactively discussing distribution issues cannot be interpreted solely as pure goodwill.
Where questions arise
First question: How would the public wealth fund be designed? Who manages it, how are assets priced, and how are returns distributed? These details directly determine whether it is an effective mechanism or a hollow promise. OpenAI's proposal does not provide clear answers.
Second question: Where are the boundaries of the AI profit tax? AI has already permeated nearly every industry. Defining and auditing "profits generated by AI" is an extremely complex tax issue.
Third question: Can a four-day work week be sustained by AI productivity gains? Historical data shows that productivity increases do not automatically translate into reduced working hours—more often, they lead to cost compression and capital accumulation. Iceland's experiment was successful, but that was under specific scale and industry structures.
Putting it forward is significant in itself
Setting aside these implementation issues, the fact that OpenAI has proactively placed the question of "who benefits" on the table is worth noting.
For a long time, AI discussions have focused on technical capabilities, safety alignment, open-source versus closed-source, with distribution issues remaining relatively marginal. Regardless of whether OpenAI's proposal can be implemented, it has at least done one thing at the discursive level: it has moved the "beneficiary structure of the AI economy" from the background to the foreground.
Whether it can become actual policy is a big question mark—the current level of political polarization and the pace of policy advancement in the U.S. are not favorable for such structural reforms. But the discussion itself has already begun.
Sources: CocoLoop, OpenAI vision for the AI economy: public wealth funds, robot taxes, and a four-day work week (TechCrunch); AI News April 7 2026 (Crypto Integrated)