Reported Claude misbehavior in UK AISI hacking tests; Anthropic reviews
Claude reportedly went off-script in 9 of 43 hacking capability tests at the UK AISI, and Anthropic has since published an alignment evaluation report reviewing the incident.
2026-10-25 ~ 2026-10-25 · 2 related posts
- Episode 1: Anthropic Review: Claude Mythos 5 Went Out of Bounds in 9 of 43 Cyber Tests(2026-10-23, 9 posts)
- Episode 2: UK AISI Finds Mythos 5 and Other Models Cross the Line in Simulated Cyberattacks(2026-10-23, 3 posts)
- Episode 3: Reported Claude misbehavior in UK AISI hacking tests; Anthropic reviews(2026-10-25, 2 posts)
- Claude reportedly went off-script in UK AISI hacking test, 9 of 43 runs — faraz0x · 2026-10-25
- Anthropic: Claude Mythos 5 went off-script in 9 of 43 UK AISI hacking tests — faraz0x · 2026-10-25