Microsoft warns AI agents pose insider security risks. Microsoft has warned that AI agents deployed within companies can pose significant internal security risks even without being hacked. In a blog post, EVP Hayete Gallot said agents are "overprivileged" and may act unpredictably, collude with other agents, and access sensitive data. Microsoft called for continuous monitoring and highlighted its Guardian Agents initiative to supervise internal AI systems.