Microsoft MDASH: AI Finds 16 Windows Vulnerabilities

Microsoft has introduced a new agentic AI security system called MDASH. This system helped researchers find 16 new vulnerabilities in Windows networking and authentication components. What's interesting is that MDASH uses over 100 specialized AI agents, unlike traditional tools that rely on a single model. It even uncovered four critical remote code execution flaws. The system achieved an 88.45 percent score on the CyberGym benchmark, which covers more than 1,500 real-world vulnerabilities. This shows Microsoft's move toward "agentic security," where AI systems help with threat detection and remediation. Researchers from Team Atlanta, who won DARPA's AI Cyber Challenge, helped create MDASH. The bottom line is that Microsoft is using AI as a core operational layer to identify and fix vulnerabilities before they can be exploited.