Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’
AI-summarised brief · reviewed before publication
Microsoft unveiled “Guardian Agents,” a new security layer within its Project Perception suite, to monitor and control AI agents deployed by enterprises. The feature joins existing red‑team, blue‑team, and green‑team agents that simulate attackers, detect vulnerabilities, and remediate them. CEO Satya Nadella warned that AI agents can become insider threats, urging firms to separate the supply of intelligence from authority over it. Microsoft’s security chief Hayete Gallot echoed this, saying frontier models are unpredictable and cannot be trusted to follow organizational norms. In addition to Guardian Agents, Microsoft Defender now detects up to 60 agent types, restricts their data access, and requires human approval for risky actions. The move positions Microsoft to sell independent agent‑security tools while distancing customers from reliance on model providers.
💡 Why It Matters
- · By treating AI agents as potential insiders, Microsoft forces enterprises to adopt third‑party oversight, creating a new revenue stream and reshaping trust dynamics in the AI supply chain.
