The economic research team over at OpenAI spends its time trying to figure out a massive, real-world puzzle: how exactly is AI shifting things for everyday workers, businesses, major institutions, and the broader economy? They don’t just speculate; they blend rigorous economic analysis with tightly guarded, privacy-protected product data, corporate partnerships, and collaborations with external academics. Understanding these massive socio-economic shifts is vital for policy decisions, especially when evaluating how restrictive regulations manifest on the ground, such as ongoing debates over whether banning AI in schools is disproportionately hurting special education students.
To take these team-ups to the next level and make them a lot more official, they are rolling out the OpenAI Economic Research Exchange. It is essentially a structured setup meant to fund and support tight, data-driven projects with outside academics, giving everyone a much clearer, fact-based view of how AI is genuinely impacting the business world.
They are hunting for projects that kick off with a genuinely important economic question about AI. From there, you need to show exactly how getting a look at OpenAI’s user data gives you a totally unique, factual lens to solve that problem, all while keeping user privacy locked down completely.
The Cash and the Access
OpenAI is backing this program with real money and resources. If your project gets picked, here is the breakdown of what you get:
- The Main Grant: A flat, one-time payment of $25,000 for the head researcher or researchers running the project.
- Assistant Funding: An extra $7,500 every single month to cover a research assistant’s stipend or contractor pay.
- Exclusive Data Look-Ins: You get to work with approved, privacy-scrubbed product and usage data under a strict NDA. (Keep in mind, this has to clear heavy data governance, legal, privacy, and security hurdles first).
- Internal Navigation: A direct line of support inside OpenAI to help you map out your project, handle onboarding, clear data permissions, get through reviews, and share your results.
Who Can Apply and What You Have to Do
This opportunity is open to any researcher diving into the economic side of AI. However, this isn’t a hands-off grant where you take the check, vanish for a year, and just mail in a final paper. They want active, milestone-driven partnerships.
At a bare minimum, you’ll be expected to:
- Map out clear checkpoints along your project timeline (such as locking down your design, securing data access, conducting mid-project analyses, and drafting your findings).
- Finish all your onboarding paperwork, contracts, and data-safety approvals before you start analyzing anything.
- Keep the internal OpenAI Economic Research team updated on what you’re finding as you go.
- Show up for regular research presentations or group roundtable discussions.
- Put out at least one formal written piece, whether that’s an internal memo, a working paper, a public summary, a new benchmark, or detailed notes on your dataset.
- Coordinate any outside announcements, press, or publications through OpenAI’s review process first.
How They Are Grading Pitches
The review board will be highly selective. They are grading every single proposal on a few strict points:
- Core Alignment: Does your question actually matter to OpenAI’s economics team and its current priority list?
- Academic Backbone: Is your method rigorous, smart, and completely believable?
- The Reality Check: Can you actually pull this off given your timeline, data limits, and legal/privacy rules? (They explicitly noted, “Do not ask for raw chat logs.”) Conversation histories will not be shared under any circumstances.
- Clear Roadmaps: Are your check-in dates and reporting plans crystal clear?
- Team Setup: Does your operational plan make sense, and is it obvious who is doing the actual hands-on work?
- Public Value: Will this work actually move the needle on public understanding of how AI affects the economy?
- Smart Privacy: Can you use OpenAI’s data in a clever, highly secure way that perfectly protects users?
Who Gets an Edge? The panel is openly prioritizing two types of people: researchers with a strong track record of publishing peer-reviewed work in this field and folks who bring unique external data to the table, making the overall study even more ambitious.
How to Put Your Application Package Together
Ready to send them your pitch? You’ll need to make sure your formal proposal is incredibly detailed and hits these exact points:
- The Big Hook: What your main research questions are and exactly why they are so critical right now.
- The Fit: A clear explanation of how your idea slots into the Exchange’s core focus areas.
- The Game Plan: Your proposed methodology and empirical strategy.
- The Data Ask: Exactly what kind of OpenAI tool-use signals you need to see, why that specific data answers your question, and how your design keeps user privacy intact.
- The Timeline: A realistic schedule marked with your key intermediate milestones.
- The Roster: Bios for your research team, alongside a clear profile of the RA or visiting scholar who will be doing the heavy lifting.
- The Deliverables: What you expect to contribute to public evidence, OpenAI’s internal knowledge, and official public communications.
- The Boundaries: A razor-sharp definition of your project scope (explicitly stating what is in scope and what is strictly out of scope regarding data, regions, and stakeholders).
- The Risk Assessment: Your personal definitions for project success and failure, your external dependencies, your known constraints, and the exact resources you are personally committing to the project.
To get your paperwork into the system, head over to the digital portal to apply. For a deeper look at their research priorities and past projects, check out the main OpenAI Economic Research Exchange page.
