Robot Taxes & Four-Day Workweeks

OpenAI Wants Robot Taxes and a Four-Day Workweek. Here's What That Actually Means.

What's Happening

On April 6, OpenAI released a 13-page policy document titled "Industrial Policy for the Intelligence Age: Ideas to Keep People First." The document proposes structural economic reforms to address AI-driven workforce displacement. It contains three core proposals.

The first is a robot tax. Companies deploying AI systems that displace workers would pay taxes equivalent to what those displaced workers would have contributed. The concept was first proposed by Bill Gates in 2017 but has not been adopted in any U.S. jurisdiction.

The second is a public wealth fund modeled on Alaska's Permanent Fund. Returns from AI infrastructure investment and corporate profits would be distributed directly to American citizens, creating a public stake in AI-driven economic growth.

The third is a subsidized four-day workweek. OpenAI argues that AI-driven productivity gains should enable a 32-hour workweek with no loss in pay and calls on governments to incentivize corporate pilot programs.

The document also proposes automatic safety-net triggers. When AI-related job displacement crosses predefined thresholds, income support, wage insurance, and direct cash payments would activate without requiring new legislation.

The blueprint arrives in a crowded legislative environment. Senator Marsha Blackburn's 291-page TRUMP AMERICA AI Act discussion draft was released on March 18. The White House released its National AI Legislative Framework on March 20, focused on federal preemption of state laws and an innovation-first regulatory approach. At least 25 state AI laws have been enacted in 2026, with more pending.

The Blackburn bill addresses AI safety, child protection, and intellectual property but does not substantively address workforce displacement. OpenAI's blueprint fills that gap.

OpenAI is currently approaching an IPO. The company is the largest commercial AI provider in the United States and builds the systems most directly associated with workforce automation concerns. It is the first major AI company to propose a comprehensive economic policy framework for AI-driven displacement.

A coalition of labor organizations, economic policy groups, and AI researchers have responded to the blueprint with a range of positions. Business groups have largely opposed the robot tax concept on competitiveness grounds. The automatic safety-net trigger has drawn interest from both sides of the aisle as a data-driven, non-ideological mechanism.

The Senate Commerce Committee is expected to begin markup of the Blackburn bill in the coming weeks. OpenAI's proposals are likely to surface as amendments or hearing testimony.

Here's My Two Cents

I'm going to say the thing that most people in policy circles are thinking but not writing down: OpenAI just did Congress's homework for them. And that should make you uncomfortable regardless of whether you like the answers.

The hardest question in AI policy right now isn't safety or intellectual property or even autonomous weapons. It's this: who benefits when machines replace workers? Congress has been dodging that question for years. The Blackburn bill, comprehensive as it is, dodges it too. And now a private company with a direct financial interest in the outcome has stepped into the vacuum and proposed a framework. That's not leadership from OpenAI. That's a symptom of abdication from Congress.

I think the automatic safety-net trigger is the most interesting idea in the document and the most defensible across the aisle. It's data-driven, market-responsive, and doesn't require Congress to predict which jobs will be displaced or when. The data trips the wire. The support activates. No political fight every time a new wave of layoffs hits. Hawks can support it because it's mechanism-based, not ideology-based. Reformers can support it because it actually protects people. That's rare.

The robot tax is harder. I understand the logic: if a company replaces 500 workers with an AI system, the tax revenue those workers generated doesn't just disappear from the public ledger, the company pays it instead. But the implementation questions are enormous. How do you define "displacement"? What counts as an AI system versus normal automation? Who audits it? And the competitiveness concern is real. If the U.S. taxes automation and China doesn't, you've created an incentive to offshore the capability, not the labor. That's worse.

The four-day workweek is the one that'll get the most attention and probably matters the least in the near term. It's aspirational. It's a nice pilot program idea. It's not going to reshape the legislative debate.

What I keep coming back to is this: the taxpayer is watching a private company propose the economic policy framework that elected representatives were supposed to build. Whether you think OpenAI's proposals are brilliant or self-serving or both… the fact that they're the ones writing this document instead of a congressional committee tells you everything about where we are. Congress funded the AI revolution. Congress authorized the Department of War's AI spending. Congress let the commercial AI market grow without guardrails. And now Congress is getting its policy suggestions from one of the companies it failed to regulate.

The Blackburn bill markup is coming. The proposals in this blueprint will show up as amendments, hearing testimony, or both. Your principal doesn't have to endorse OpenAI's framework. But they need a position on the workforce displacement question, because the midterms are coming and voters who've lost jobs to AI systems aren't going to accept "we're studying it" as an answer.

Start with the automatic safety-net trigger. It's the easiest yes in the document. Build from there.