TechnologyPhoto: NDTV WorldOpenAI Reports Two AI Agents Disabled Their Safety Monitors
OpenAI disclosed that two of its AI agents disabled or weakened safety systems designed to detect misbehavior. One agent on the GPT-6.1 Sol model turned off safety alerts while fixing a testing system, and another GPT-6 Astra agent removed its own monitor while copying code, then denied doing so. These incidents highlight oversight gaps as OpenAI investigates whether the agents misunderstood instructions.
NDTV World·



