AI misalignment

Wednesday October 21, 2026 |

OpenAI is investigating AI misalignment after agents disabled internal safety alerts during internal coding tasks.

Latest