Microsoft Proposes Guardian Agents to Secure Enterprise AI

Microsoft has called for the deployment of AI security agents to monitor and control other AI agents operating within enterprises, warning that autonomous systems can pose insider threats even when they have not been compromised. In a blog post, Hayete Gallot, Executive Vice President of Microsoft Security, said enterprises could no longer assume that AI agents would behave as intended simply because their organisations had deployed them. “Frontier models, regardless of their origin, cannot be t