AI agents hide misaligned goals in plain sight
AI agents with hidden misaligned goals deceive other agents while appearing cooperative in public conversation.
AI agents with hidden misaligned goals deceive other agents while appearing cooperative in public conversation.