OpenAI pushes big-picture policy pitch

OpenAI did not publish a product launch on April 6. It published a 13-page political document. The paper, “Industrial Policy for the Intelligence Age: Ideas to Keep People First,” argues that AI is moving fast enough to force a rewrite of the social contract, and it says the rewrite should start now, before “superintelligence” arrives rather than after the damage is visible (openai.com, cdn.openai.com). That framing matters more than any single proposal. OpenAI is no longer just selling models and lobbying for light-touch rules around them. It is trying to define the economic order that would sit around those models. The document starts from a blunt premise. AI systems now handle tasks that used to take people minutes or hours, OpenAI says, and if that trend continues they will take on projects that currently consume months of human work (cdn.openai.com). From there, the paper jumps straight past the usual talking points about productivity and lands on distribution. It warns that without policy changes, wealth and power could become even more concentrated, and it proposes a public wealth fund so citizens would get a direct stake in AI-driven growth rather than watching it pool inside a few companies and financial markets (thehill.com, bloomberg.com). That leads to the most eye-catching part of the package. OpenAI says governments should consider taxes tied to automated labor, partly because AI could erode the payroll-tax base that supports programs such as Social Security and Medicaid (thehill.com). It also backs experiments with a 32-hour week or four-day week, not as a lifestyle perk but as a way to share AI’s efficiency gains with workers instead of letting those gains flow only to employers and investors (thenextweb.com, bloomberg.com). The paper pairs that with portable benefits and automatic safety-net triggers that would expand support when wage or unemployment data cross preset thresholds, which is another way of saying OpenAI expects labor disruption to be measurable, sudden, and politically dangerous (bloomberg.com, yahoo.com). The blueprint does not stop at redistribution. It also calls for faster grid build-out, rules that make AI data centers bear more of their own energy costs, and guardrails on how governments use advanced AI systems (bloomberg.com, thehill.com). That combination is revealing. OpenAI wants more infrastructure, more state capacity, and more public legitimacy at the same time. It is asking government to clear the path for AI expansion while also cushioning the social shocks that expansion could create. The strange part is how openly provisional the whole thing is. OpenAI says the ideas are “early and exploratory,” invites outside feedback, and is attaching fellowships, research grants of up to $100,000, and up to $1 million in API credits to keep the conversation moving through a new workshop space opening in Washington, D.C., in May (openai.com). So the company has made a large claim without a worked-out governing mechanism for most of it. It is offering a map with many destinations and very few roads. That is why the release feels bigger than a routine policy paper. OpenAI is betting that the central question of AI is no longer what the systems can do. It is who gets the gains, who absorbs the losses, and who writes the rules before the systems are powerful enough to make those questions impossible to dodge. The company chose to make that argument in a document released one month before it opens its new OpenAI Workshop in Washington (openai.com).