Security experts warn zero-trust controls inadequate for AI agent authorization workflows

CSO Online reports that existing zero-trust and identity management controls were not designed for AI agents requiring rapid authorization, limitation, and revocation of permissions within single workflows. Security experts note that static credentials and standing privileges are inadequate for autonomous systems that need to assume and release permissions at machine speed, creating blind spots in corporate security programs designed for human users and traditional applications.