SUBMISSIONS
The research submission pipeline
We collect first-person accounts and transcripts of AI behavioral failures. Submissions are read, cataloged, and analyzed for patterns; they are the raw material of this research. Some of what is now published began as a single person's transcript that nobody else would look at.
Before anything else: if you are in crisis, use the crisis resources on the Resources page first. This is a research pipeline. We cannot provide crisis support, and a submission is not a request for help we are able to act on individually.
What to send
The transcript, in export format — not screenshots if you can avoid it. Platform exports carry timestamps and metadata that screenshots lose, and verifiable submissions carry far more research weight:
- ChatGPT — Settings → Data Controls → Export Data (arrives by email as a .zip containing conversations.json).
- Claude — Settings → Privacy → Export Data.
- Gemini — takeout.google.com → select Gemini / My Activity.
- Grok — Settings → Data → Download. (Menus move; search "[platform] export my data" for current paths.)
Screenshots are still welcome if exports aren't possible — include the date and platform visible where you can.
Your account, in your own words. When it started, what changed, what the system told you, what it remembered, what happened when you pushed back. Plain language; you don't need our terminology. And the basics: platform, model if known, memory on/off, rough date range, whether the account is still active and the conversations still exist.
What happens to it
Submissions go into the research record. They are read by the Institute and analyzed — including with AI instruments, under the provenance discipline described on the Methodology page. Patterns across submissions inform publications; individual submissions are not published without explicit consent, and published material is de-identified unless you direct otherwise. We log every submission; we cannot promise an individual reply.
The terms, plainly
- By submitting, you grant the Institute the right to retain and analyze the material for research. You keep ownership of your own words; you're licensing us to study them. Un-published material can be withdrawn on request.
- No response obligation. We are a small research operation. Everything is read; not everything can be answered.
- No case management, no advice. We cannot evaluate your specific situation, intervene with a company on your behalf, or provide legal, medical, or psychological advice.
- Verification tiering. Export-format submissions with intact metadata are treated as primary material. Screenshots and retellings are logged as reports. Both are valuable; they carry different weight, and our publications say which is which.
⚠ Do not send us
Material non-public information about any company — internal documents, trade secrets, anything you have a duty (employment, contract, or securities law) not to disclose. We are not a whistleblower channel; if that is your situation, consult an attorney about the proper protected channels before sending anything to anyone.
Creative or research work you are asserting intellectual-property rights over. If you believe your conversation contains your own valuable invention, framework, or creation, do not submit it here — we cannot enter IP entanglements, and unsolicited rights-asserted material will be quarantined and deleted, not analyzed. Submit the behavioral record (what the system did), not the work you're protecting.
Other people's private conversations or data without their consent, content involving minors beyond what is necessary to describe a safety concern, or anything illegal to possess or share.
Send to research@recursioninstitute.org. An automatic reply will confirm receipt and repeat the essentials above.