Microsoft adds Guardian Agent to Project Perception security system
Microsoft says the Guardian Agent monitors internal AI agents around the clock and acts before their actions are executed, VP Yvette Gallo wrote in a company blog post.
Microsoft has extended its Project Perception security system to bring internal AI agents under control through a new component called the Guardian Agent, the company said in an official blog post on October 19. The Guardian Agent is designed to watch over agent actions 24 hours a day and to check them before they are executed.
Project Perception locates agents operating inside a company using signals from identity, data, cloud, endpoint and network sensors. Each agent it finds is given an identity, along with policies and permissions, and is then brought under management.
The system also calculates the risk level of agent actions in real time. When that risk rises, the Guardian Agent responds, according to Microsoft.
Traceable records for security teams
Microsoft Vice President Yvette Gallo presented the approach in the blog post. Security staff can trace the records to establish why a risk arose and then improve the system, the company said.
The announcement was reported by digitaltoday.co.kr.
Quick answers
What is Microsoft's Guardian Agent?
It is a component of Microsoft's Project Perception security system that monitors internal AI agents 24 hours a day and checks their actions before execution.
How does Project Perception identify AI agents?
It uses signals from identity, data, cloud, endpoint and network sensors, then assigns each identified agent an identity with policies and permissions.