Microsoft’s Project Perception brings coordinated AI agents into security operations, raising new questions about permissions, oversight, accuracy, and deployment risk.
