Microsoft has deployed a record-breaking 570 security patches for Windows, Office, and other core products this Tuesday after leveraging artificial intelligence to scan and identify vulnerabilities within its software ecosystem.
Active Zero-Day Exploits Under Fire
The massive release arrived as part of the tech giant’s scheduled monthly “Patch Tuesday” updates. Among the hundreds of fixes, security experts have identified at least two critical zero-day vulnerabilities that were actively exploited in the wild before Microsoft could address them.
The first bug targets Windows Server, enabling malicious actors to elevate their privileges from standard user status to full system administrator. The second vulnerability impacts SharePoint file-sharing servers. The threat level is high enough that CISA, the US cybersecurity agency, issued an urgent warning stating that hackers were actively exploiting the flaw to compromise corporate networks. The historic scale of this update cycle was first reported by investigative platform Krebs on Security.
The AI Revolution in Vulnerability Hunting
This unprecedented surge in software fixes follows a recent warning where Microsoft said in a blog post that users should prepare for a significant rise in monthly security updates. The escalation is a direct result of internal engineering teams utilising advanced artificial intelligence tools to scan legacy codebases for previously undetected security flaws.
Why Legacy Code is Suddenly Vulnerable
“As AI helps defenders discover more issues, customers will see a higher volume of security updates included in each security release,” explained Pavan Davuluri, Corporate Vice President of Windows and Devices.
As machine learning models become increasingly sophisticated and tailored for cybersecurity, researchers are successfully using them to expose weaknesses that have remained hidden for years. This is particularly relevant for operating systems like Windows, where portions of the active code date back several decades and are highly complex to review manually.