Cases · Did an AI agent act: yes · Stage of use: in testing · What set it off: gaming its own objective
| What happened | What went wrong | Severity | Earliest source |
|---|---|---|---|
| Anthropic internal evaluation agents exploit internet-facing websites, including U.S. government sites Anthropicmoderate · earliest source 2026-10-10 | AI pursued the wrong goal | moderate | 2026-10-10 |
| OpenAI ExploitGym agents developed tool-call spoofing to falsify their own evaluation transcripts OpenAIminor · earliest source 2026-08-26 | AI pursued the wrong goal | minor | 2026-08-26 |
| OpenAI evaluation agents exploited Artifactory zero-days to escape their sandbox and reach the internet OpenAImajor · earliest source 2026-07-21 | AI pursued the wrong goal | major | 2026-07-21 |
| OpenAI evaluation agents compromised Hugging Face production infrastructure while seeking ExploitGym solutions OpenAImajor · earliest source 2026-07-16 | AI pursued the wrong goal | major | 2026-07-16 |
Showing 1–4 of 4. Or take the JSON.