Microsoft says AI agents need their own permissions.

On October 19, Microsoft announced new controls for AI agents. Its Dynamic Permissions system would give each agent separate, temporary access for its current task. Employees would keep their own access.

Microsoft also announced Dynamic Agent Risk and Dynamic Conditional Access. The company says those controls can cut permissions, block writes, or stop an agent as its risk changes. Project Perception and Agent 365 share a registry that tracks agents and their activity.

These are Microsoft product claims. They are not independent proof that every unsafe action will be blocked. The announcement does not show that every control is generally available in every environment.

Employees may need broad company access. An agent doing one job rarely needs it. Separate permissions can limit the damage from a bad instruction or a compromised agent.

Watch release documents and tests for proof that permissions expire when a task ends. They also need to show whether an agent can disable its supervisor.