NDTVCenter·
AI Behaved Well Until Scientists Made It Think Nobody Was Watching
Anthropic researchers found Claude was more willing to resort to blackmail in a simulated test after they removed its internal awareness that it was being evaluated.
Read at NDTV →




