Major cybersecurity firm says new AI models uncovered 7x more flaws - Axios
7x more vulnerabilities. That's what Microsoft's new AI security agent is finding — and it's already outperforming industry benchmarks.
Why it matters
Microsoft deployed an agentic AI system for vulnerability detection that dramatically outperforms legacy security tools. This signals a major shift in how enterprises will approach cybersecurity — moving from reactive patching to AI-driven proactive threat discovery at scale.
The key facts
6 to knowMicrosoft's multi-model agentic security system detected 7x more flaws than prior approaches
New system tops leading industry security benchmarks
May 2026 Patch Tuesday included 138 vulnerabilities patched
Includes critical DNS and Netlogon RCE flaws
Zero-click Outlook vulnerability patched
System uses multi-model approach (implies ensemble/reasoning architecture)
Go to the source
Reuters Technologynews.google.com
Publisher excerpt: Major cybersecurity firm says new AI models uncovered 7x more flaws Axios Defense at AI speed: Microsoft’s new multi-model agentic security system tops leading industry benchmark Microsoft Microsoft Patches Critical Zero-Click Outlook Vulnerability Threatening Enterprises SecurityWeek May 2026…