OpenAI calls for robot taxes and a public wealth fund to cushion AI job losses

OpenAI has published a policy blueprint calling for robot taxes, a public wealth fund, and trials of a four-day workweek as part of a broad set of proposals designed to cushion the economic disruption expected from artificial intelligence.

The 13-page document, titled "Industrial Policy for the Intelligence Age: Ideas to Keep People First," was released Monday. It frames the proposals as a starting point for public debate rather than a finished prescription, Axios reported, which published an interview with CEO Sam Altman alongside the document's release.

Every American would receive an ownership interest in the gains produced by artificial intelligence under one of the document's more sweeping proposals — a nationally managed public wealth fund that Axios characterized as the blueprint's most far-reaching element. Contributions from AI companies would help capitalize the fund, which is envisioned as holding stakes across both the AI sector and the wider range of industries adopting the technology.

Tax policy proposals in the document include charges tied to the use of automated workers and a restructuring of the sources of government revenue — moving the emphasis away from wages and toward investment returns and corporate profits. Underlying the tax proposals is the concern that widespread automation could erode the employment-based income streams that fund Social Security, Medicaid, and SNAP.

Workers would see AI productivity improvements translate into shorter hours rather than higher output under another proposal, which calls for government-backed experiments with 32-hour schedules that maintain current pay levels. OpenAI's chief global affairs officer Chris Lehane told Bloomberg the policy conversation around AI needs to be "as transformative" as the technology itself.

The document also envisions a data-driven mechanism that would expand government assistance without requiring new legislation each time — once measurements of AI-related job displacement cross defined limits, programs covering income support, wage insurance, and direct cash payments would activate automatically. As labor market indicators recovered, the expanded benefits would wind down on their own. Rounding out the social proposals, the blueprint argues that access to AI tools should be treated as a basic public entitlement on par with reading ability or electrical service, and that pricing must not put those tools out of reach for hourly workers, community institutions, or economically marginalized groups.

Perhaps the document's starkest moment comes when it confronts the possibility of AI systems that spread and operate beyond human control — machines that, because they can copy themselves and act independently, could not be shut down through conventional means, making pre-arranged government-level response plans essential.

Speaking to Axios, Altman framed the pace of superintelligence development as demanding a reimagining of American society's foundational agreements — a transformation he compared in ambition to the Progressive $PGR Era reforms of the early twentieth century and the New Deal responses to the Depression. Of all the risks on the horizon, Altman singled out cyber and biological threats as the dangers that concern him most in the near term. "I think that's totally possible," he said of a significant cyberattack occurring within a year. "I suspect in the next year, we will see significant threats we have to mitigate from cyber."

The backdrop to the proposals is a labor market already showing strain. White-collar payrolls have contracted for 29 consecutive months, a stretch economists describe as unprecedented outside a recession, and researchers have documented a decline in demand even for elite business school graduates. This analysis found that AI is reducing demand for white-collar workers, while the technology's positive job-creation effects remain years away.

The document offers its own definition of superintelligence — machines that surpass even the most capable humans at cognitive tasks, including situations in which those humans are working alongside AI tools. Bloomberg reported that ChatGPT's global weekly user base has grown to around 900 million people.