AI agents governed by contracts and judges
A new safety framework constrains AI agents to approved actions, enforced by separate monitor and judge agents.
A new safety framework constrains AI agents to approved actions, enforced by separate monitor and judge agents.