Microsoft Fixes Record Number of Security Flaws, Credits AI for Faster Vulnerability Detection

Microsoft has rolled out its largest-ever batch of security updates for Windows, Office, and other software products, saying that artificial intelligence is helping engineers identify software vulnerabilities more quickly.

As part of this month’s scheduled Patch Tuesday release, the technology company published fixes for 570 security vulnerabilities across its product portfolio. The monthly security update cycle, commonly referred to by researchers as “Patch Tuesday,” has long served as Microsoft’s regular release window for security patches.

Among the vulnerabilities addressed are at least two zero-day flaws, meaning they had already been exploited before Microsoft became aware of the issues. One of the bugs affects Windows Server and could allow attackers to elevate privileges from a standard user account to full system administrator access. Another impact is on Microsoft SharePoint, with the U.S. Cybersecurity and Infrastructure Security Agency (CISA) warning that threat actors were actively exploiting the flaw to compromise organisations.

Krebs on Security first reported the unusually large security release.

The record-breaking update comes just one week after Microsoft said in a company blog post that future Patch Tuesday releases would likely include significantly more security fixes than in previous years. The company attributed the increase to its growing use of AI tools that assist security teams in identifying previously undiscovered flaws throughout its software.

“As AI helps defenders discover more issues, customers will see a higher volume of security updates included in each security release,” said Windows chief Pavan Davuluri.

As AI models become increasingly capable in cybersecurity research, security professionals are using them to uncover vulnerabilities that may have remained hidden in software for years or even decades. In Microsoft’s case, portions of the Windows codebase date back many years, providing AI with a vast amount of legacy code to analyse for previously undetected security weaknesses.