Microsoft expands Project Perception to secure AI agents
According to Microsoft’s official blog, Microsoft expanded Project Perception with continuous agent risk assessment, adaptive permissions, and threat containment to treat AI models as insider risks.
SourceAnalysis
Microsoft CEO Satya Nadella announced on October 19, 2026 the expansion of Project Perception with new capabilities designed to continuously assess agent risk, adapt permissions and contain threats.
The update puts into practice the principle of treating models as insider risks with security controls positioned outside the model itself according to Satya Nadella on X and the linked Microsoft blog post.
Satya Nadella
@satyanadellaChairman and CEO at Microsoft