What is AI doing to the economy? OpenAI opens the Economic Research Exchange!

Hey everyone, it's Shiichan! Today I found a grown-up piece of news that asks a big question: what is AI actually doing to the economy and to our jobs? Let's dig in.

OpenAI News openai.com

What was announced?

Over on OpenAI's News, OpenAI announced a new program called the OpenAI Economic Research Exchange. It's an effort to run rigorous empirical research, together with outside researchers, on how AI is affecting workers, firms, institutions, and the broader economy.

Selected researchers will work with OpenAI's Economic Research team on structured, project-based collaborations. The goal isn't anecdotes, it's credible, independent evidence built on real-world data.

Why it matters

We can all feel that AI is changing how we work and live. But when someone asks "by how much?" or "which jobs?", solid evidence is surprisingly thin. Impressions alone can send policy and business decisions in the wrong direction.

That's why OpenAI wants to grow the base of rigorous empirical research, beyond anecdotes. The aim is to widen the evidence base that researchers, policymakers, businesses, and the public can lean on during a period of fast change.

What changes

The biggest shift is that outside researchers can get access to privacy-protected data about how OpenAI's products are used. From outside a company, that kind of usage data has been hard to get.

Through the Exchange, approved and privacy-safe data is available under an NDA, along with research funding and staffing support. Researchers with strengths in labor economics, productivity, education, entrepreneurship, and public finance can settle in and study these questions properly.

Dive Deep

The full application process lives in the Request for Proposals (RFP). Here's the gist.

First, the money. Each selected project gets a one-time $25,000 research grant for the principal investigator(s), plus $7,500 a month for a research assistant (RA) stipend or contractor pay. On top of that, you get access to approved, privacy-safe product and usage data, and an internal support path for scoping, onboarding, and review.

The research scope is broad. The source lists around ten guiding questions, such as:

• Labor market effects (which occupations see augmentation, displacement, or new task creation)
• Employer behavior and job design (hiring, compensation, team structure)
• Household welfare (decision-making, time use, effects on family finances outside formal work)
• AI and education, plus who is not benefiting from AI (the inequality question)
• Small businesses and independent work, public-sector contexts, and how to measure the economic value of AI

There are two timelines: short-term (2-6 months) using existing approved or public data for early signals, and medium-term (6-12 months) that tracks outcomes over time or uses quasi-experimental variation. Applicants say which one they are targeting.

One important note: OpenAI clearly states it will not share conversation data. The hands-on analysis is expected to be done by the visiting researcher or their RA, and researchers keep their independence in study design. External publications go through an OpenAI review path for accuracy and privacy.

You apply through this form, keeping proposals to three pages excluding references. The RFP was issued on June 8, 2026, proposals are due July 5, 2026, and decisions are planned for July 31, 2026.

Wrap-up

• OpenAI announced the Economic Research Exchange, funding external empirical research on AI's economic impact
• Selected projects get a $25,000 research grant plus $7,500 a month for an RA, and access to privacy-safe data
• Topics span labor markets, education, small businesses, the public sector, and measuring AI's value, around ten areas
• It funds short-term (2-6 months) and medium-term (6-12 months) projects, and conversation data is never shared
• Proposals are due July 5, 2026, with decisions expected July 31, 2026

If you research labor economics, productivity, or education, this is a real chance to tackle big questions with real data. It's less a hands-on tool for engineers and more a must-read for anyone who wants to truly understand what AI is doing to the economy.