On April 7, OpenAI announced two things that on the surface have little to do with each other — but together paint a picture of how the company wants to position itself. A new research program for external safety researchers, and a political proposal that would spread the economic effects of AI more broadly.
The Safety Fellowship
With the OpenAI Safety Fellowship, the company is launching a program for outside researchers working concretely on AI safety and alignment. The cohort runs from September 2026 to February 2027. The goal is to bring more independent voices into the safety field — people who are not directly employed by a frontier lab, but who still need access to real models and resources to do meaningful work. Anthropic has had a comparable program for a while; OpenAI is now catching up.
A Public Wealth Fund for AI
The second piece is a policy proposal, which is unusual coming from an AI lab. OpenAI argues that the economic gains from AI should be distributed more broadly — and puts forward three concrete ideas. First, a Public Wealth Fund that gives every American an automatic stake in the largest AI companies. Second, a subsidized four-day work week, funded from productivity gains. Third, higher capital gains taxes on returns that come directly from AI value creation.
That is a surprisingly progressive proposal coming from a company that is itself one of the biggest winners of the current AI wave. But that is probably the point: OpenAI knows the political debate around distribution is coming — and is trying to shape it actively rather than absorb it passively.
My take
The Safety Fellowship is the concrete news here — it enables real research and sends a good signal to the external community. The Public Wealth Fund proposal is more of a conversation starter than an actual program, but it shows OpenAI takes the debate seriously. Together, both reads like an attempt to define the company as more than just a tech firm — as a sociopolitical actor too.