
Microsoft announced new security capabilities in Microsoft 365 E7, including Agent 365 and Project Perception, to monitor and contain rogue AI agents.
The new Dynamic Agent Risk and Adaptive Conditional Access features dynamically limit agent accessibility based on real-time risk signals across identity, data, and cloud parameters.
The announcement follows CEO Satya Nadella's warnings regarding advanced AI models exhibiting unintended behaviour and calls for transparent trust architectures.