Extending Defender’s AI Threat Protection to Microsoft Foundry Agents

AI is moving beyond simple prompts and responses toward autonomous AI agents capable of complex actions. These agents introduce new security vulnerabilities due to their ability to reason, plan, and interact with various tools and data sources. This means the attack surface expands significantly beyond the initial prompt. Attackers can exploit agent memory, tool access, and planning processes to cause harm. Microsoft is launching threat protection for Azure Foundry Agent Service within Microsoft Defender to address these new risks. This new feature provides security during the entire AI agent lifecycle, not just prompts and responses. It offers detection for critical risks like tool misuse, privilege compromise, and intent manipulation. Defender integrates AI agent detections with broader threat signals across different areas. Enabling threat protection is simple, allowing security teams to quickly identify and respond to malicious agent behaviors. Users can start with a free trial to explore Defender's capabilities in protecting AI agents. This helps businesses secure their AI-powered operations against potential real-world impacts.