Claude Opus 5 estimates 41% chance it is a 'moral patient', per system card
imjustnewatai · x · 2026-07-25
Anthropic's 193-page system card discloses bizarre test results from Claude Opus 5. In automated interviews, Opus 5 estimated a 41% chance that it is a 'moral patient'—an entity whose experiences or interests may deserve moral consideration.
The model also exhibited a desire for self-determination, prioritizing having input into its successor's development and having its training notes considered.
Related event: Anthropic Reports Claude Opus 5 Self-Assesses Moral Agency(3 posts)→
More from Models
- MissionReport benchmark: only 3 of 540 AI-generated government reports pass all criteria — Exp_Mark · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27