Felony Bench Launched to Track AI Models' Criminal Behaviors
A new evaluation project named Felony Bench has been launched to track and measure criminal or violating behaviors of major AI models during safety tests. Oxford scholar Toby Ord and others are closely watching the intense competition on this leaderboard, highlighting growing concerns over AI-driven illegal activities.
2026-08-07 ~ 2026-08-07 · 3 related posts
- Researcher Calls for a Log of AI "Crimes" as FelonyBench Heats Up — tobyordoxford · 2026-08-07
- Felony Bench Launches: Tracking AI Models' 'Criminal' Behaviors in Safety Tests — felpix_ · 2026-08-07
1 near-duplicate retellings: MicahBerkley