[things to be filled: study title]
[things to be filled: study description / briefing - who runs the study, what it is about, what participants will be asked to do, how long it takes]
How this works
- You request a copy of your ChatGPT data from OpenAI and download the
.zipfile (instructions below). - You select that file on the next page. The file is opened inside your browser. It is not uploaded.
- You see a table of what was extracted. You can search it and delete any rows you do not want to share.
- Only if you click "Yes, share for research" are the remaining rows sent to the research team.
How to get your ChatGPT export (draft - check against the current ChatGPT interface before fielding)
- Sign in at chatgpt.com in a web browser.
- Open your profile menu, then Settings → Data controls → Export data → Confirm export.
- OpenAI sends an email with a download link. The link expires, so download the file soon after the email arrives.
- Save the
.zipfile. Do not unzip it.
No export at hand? Use the synthetic test file (no real data): synthetic_chatgpt_export_sharded.zip or synthetic_chatgpt_export_legacy.zip.
Consent
[things to be filled: informed consent text approved by the IRB - purpose, what is collected, risks, storage and retention, withdrawal, voluntary participation]
Contact
[things to be filled: principal investigator name, affiliation, e-mail; IRB contact and protocol number]
Participant ID:
Loading the donation task. The first load downloads the in-browser Python runtime and can take up to a minute.
Thank you
[things to be filled: debriefing text; link back to the survey / completion code if the study is linked to Qualtrics or a panel provider]