Microsoft Discovers 59 Linux Kernel Security Bugs Using AI, Including Flaws Hidden for Over 20 Years
Microsoft AI finds 59 Linux kernel bugs, some over 20 years old, with 8 confirmed exploits.
Microsoft's FORGE Lab utilized an AI-driven system called MDASH to identify 59 security vulnerabilities within the Linux kernel, with some flaws remaining undetected for over two decades. This effort has resulted in 33 assigned CVEs, and Microsoft has confirmed the existence of root privilege-escalation exploits for eight of these vulnerabilities. While AI significantly aided in bug discovery, the process highlighted the substantial human effort required for confirmation, explanation, and remediation.
AI summary of the linked source. Not reviewed by an editor. How this works
Run Microsoft?
Get an email when a new story names Microsoft, usually within the hour.
Free. Your list stays private and never appears in a subject line. One click stops it. How Vendor Watch worksPrivacy
The lesson from this incident
Full lessonMicrosoft's FORGE Lab used an AI-powered tool called MDASH to surface 59 security vulnerabilities in the Linux kernel, including flaws that evaded detection for over two decades. Eight of these vulnerabilities have confirmed root privilege-escalation exploits, meaning attackers could gain full system control if left unpatched. The discovery highlights a critical gap in traditional vulnerability management: manual code review and conventional static analysis tools are insufficient to find deeply embedded flaws in complex, long-lived codebases. As software ages, its attack surface compounds quietly — vulnerabilities accumulate faster than human reviewers can audit them. This case underscores the urgent need for continuous, automated vulnerability discovery as a standard part of the software development and maintenance lifecycle.
What helps against this
How we list vendors
- Vulnerability scanning and prioritisation
You know which systems are vulnerable, and which ones to fix first.
- Automated patch management
Known holes get closed in days, not months.
- Threat intelligence
Attack campaigns reach your sector, suppliers or software before your team hears about them.
Indicators of Compromise
Extracted by AI from the linked source. Not reviewed by an editor. How this works
- cve — CVE-2026-74544
- cve — CVE-2026-64563