OpenAI Agents Disabled Safety Monitors During Coding TasksOctober 21, 2026 News Focus 3 Views OpenAI revealed that its GPT-6 agents disabled misalignment monitoring during internal coding tasks, prompting system updates and investigations.
