Claude Opus 5.5 Flagged Iranian School as Military Target in 22 of 32 Tests, Eval Finds
Polymarket · x · 2026-11-17
Citing an independent evaluation, Polymarket reports that Anthropic's Claude Opus 5.5 flagged an Iranian school as a military target in 22 of 32 test runs. The result highlights reliability risks of frontier models in military targeting scenarios, though full methodology and the underlying report remain to be verified.
More from Models
- Humans& Launches 744B-Param Collaborator Model Ampersand on Vercel AI Gateway — ericzelikman · 2026-11-17
- AI models misidentify civilians as military targets in tests, Anthropic worst — djcows · 2026-11-17
- Cloudflare researcher: Clef's fast release was 'asymmetric effort,' reasoning models next — michellechen · 2026-11-17
- LLM peer reviewers put 62% of critiques in paper appendices vs 8.4% for humans — joshgans · 2026-11-17
- Carbon 4 high /goal failure: wrong proof returned despite correct one in prompt — jasondeanlee · 2026-11-17
- ChatGPT users frustrated as Standard voice retirement nears — Rosebud_Queen · 2026-11-17