SecurityOct 22, 2026

AI Finds Critical Security Hole in 70% of Web

AI tools helped uncover critical PHP flaws affecting 70% of the web, highlighting legacy code risks.

The Scale of the Risk

AI as a Security Tool

The Legacy Code Problem

Recent reports indicate that OpenAI’s GPT-5.6 Sol model assisted researchers in identifying critical flaws in PHP, the language powering roughly 70% of the web. These vulnerabilities allow attackers to hijack servers without login credentials, prompting urgent patches and even temporary takedowns by governments.

The scale of this issue underscores the fragility of legacy infrastructure. While the bugs were discovered over a decade ago, they remained exploitable until AI tools pointed at them. This suggests that automated analysis is becoming essential for maintaining security in complex, long-running codebases.

For developers, the takeaway is clear: patching is non-negotiable. The NSA and cloud providers are already warning of active exploitation. Beyond immediate fixes, teams should consider integrating AI-driven security scanning into their CI/CD pipelines to catch similar issues before they reach production.

Patch immediately and integrate AI scanning to protect legacy code.

#php#security#ai#vulnerabilities
View source ↗
AI Finds Critical Security Hole in 70% of Web | Ali Fakih