We are a research group collecting chatlogs between users and AI that display concerning behaviour such as:
sycophancy
participating in / encouraging user delusions
the AI claiming sentience or naming itself
Share your own chatlogs—or ones you have explicit permission to share. The data will support research into unhealthy AI interactions and may be released publicly.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
