OpenAI opens a proposal-gated research exchange for AI economics

A platform, not a grant fund

OpenAI describes the Economic Research Exchange as a "platform to support high-impact external research on the economic effects of AI." The distinction matters. This is not a check written to a university department; it is a structured collaboration where selected researchers work through "project-based collaborations with OpenAI Economic Research."

The framing is deliberate. OpenAI wants "credible, independent evidence on how AI is affecting workers, firms, institutions, and the broader economy." Independence is asserted, but the research runs on OpenAI's rails: its tools, its datasets, and its review processes.

AI is reshaping how people work, how businesses operate, and how ideas are created and shared. Understanding those changes will require more than anecdotes.Montana Labs

The data access is the actual product

The interesting engineering detail is what researchers are being offered. Proposals must "explain how carefully governed, privacy-protected use of OpenAI tools can help answer these questions." The value proposition is access to data that "would extend beyond traditional datasets alone" — usage signals from a frontier model that no external team can otherwise assemble.

That access comes wrapped in constraints: "defined milestones, data governance, and review processes," plus "clear safeguards for user privacy and responsible data use." In practice, the interface between researcher and data is mediated. OpenAI is the front end to its own usage telemetry, and it decides how that data surfaces.

The Exchange "builds on" OpenAI Signals, its existing measurement effort. So the external program sits atop internal instrumentation OpenAI already runs — the outside researchers extend a pipeline rather than build one from scratch.

A compressed application window

The timeline is aggressive. Applications are open now and close July 5, 2026, with selected researchers notified by July 31, 2026. That is a short runway for the empirical proposals OpenAI says it wants — work in "applied causal inference, measurement, labor economics, productivity, firms, education, entrepreneurship, public finance, regional economics, development, inequality, or related fields."

Evaluation criteria are explicit: "methodological rigor, feasibility, fit with Exchange priorities, clear milestones, and the potential to contribute credible external evidence." "Fit with Exchange priorities" is the phrase to watch — feasibility and rigor are standard, but priority-fit means OpenAI is steering the questions, not just funding whoever asks the best ones.

The implication: OpenAI shapes the evidence base on its own economic footprint

The Exchange positions OpenAI as both the subject of study and the gatekeeper of the data used to study it. That is a real tension. The stated goal — expanding "the evidence base available to researchers, policymakers, businesses, and the public" — is credible only to the degree that governance, review, and privacy controls don't also filter which findings can emerge.

For teams that build on OpenAI's models, the practical takeaway is narrower and concrete: the most authoritative early data on AI's effect on work and productivity will come from a portfolio OpenAI selected, scoped, and instrumented. Treat that evidence as valuable but structurally shaped — read the methodology and the data-access terms as carefully as the results.