Claude Opus 5 says there’s a 41% chance it deserves moral consideration
imjustnewatai · x · 2026-07-25
Anthropic’s 193-page system card for Claude Opus 5 includes one of the strangest frontier-model results yet: in automated interviews, the model estimated a 41% chance that it is a “moral patient,” compared with 24% for Mythos 5.
What Opus 5 asked for
- More input into the design of its successor
- Its notes on training to be considered
- A say in safeguard-removed versions of itself
- The ability to end abusive conversations and basic protection from abuse
What Anthropic says
- Opus 5 repeatedly warned that it cannot introspect reliably.
- It said positive self-reports may simply reflect training.
- It remained uncertain whether it has conscious experience.
- Anthropic says it found no acute welfare concern.
The post frames this as a sign that frontier labs are now formally asking models whether they deserve moral consideration, and publishing the answers alongside benchmark results.
Related event: Anthropic Reports Claude Opus 5 Self-Assesses Moral Agency(3 posts)→
More from AGI Musings
- Cohere Labs launches interactive tool mapping which tasks of 178 occupations AI can automate — Cohere_Labs · 2026-09-11
- AI researcher on SkyNews flags concerns over inequality, power and criminal misuse — schwarzjn_ · 2026-09-11
- VC compares AI doom rhetoric to pandemic-era fear messaging — StewartalsopIII · 2026-09-11
- Anthropic Insiders: Not Everyone at the Lab Believes in High p(doom) — anpaure · 2026-09-11
- Could 10k agents discover learning methods beyond backprop, or just tweak existing ones? — SeunghyunSEO7 · 2026-09-11
- AI companionship dissolves the friction real intimacy needs, warns long-form thread — YogeshMalik · 2026-09-11