Explain the event
The story comes before the technical explanation. We show what the AI did and how human decisions, permissions, and safeguards shaped the outcome.
About AIxploit
Some AI incidents deserve more than a headline.
AIxploit is an independent editorial project covering consequential AI behavior. Our incident stories reconstruct concrete events; Editorial explores research, patterns and questions that extend beyond one case.
We select cases for what readers can learn from them. Each story asks what the AI did, how people and systems shaped the outcome, what followed, and where the evidence ends.
How we work
Clear accounts. Evidence in view.
Show the evidence
Substantive claims link to public sources. We attribute investigators' findings, distinguish observation from interpretation, and keep meaningful uncertainty visible.
Keep perspective
AIxploit is selective. We give each event the scale its evidence warrants, including cases where the outcome was limited or a feared loss was reversed.
Update openly
New evidence can change a story. Material updates and corrections belong with the article so readers can see how our understanding developed.
START READINGSTORIES
Begin with the stories themselves.
Explore all storiesTake part
Help build AIxploit.
Have a well-documented AI incident to share, or interested in research, translation or editorial work? Send us a note about how you would like to help.
Write to us