OpenAI says AI helped uncover critical PHP flaws

OpenAI says researchers using GPT-5.6 Sol helped uncover two critical PHP vulnerabilities, rated 9.2 in severity, and demonstrated working exploits. The company says fixes were then coordinated.

Why it matters: AI-assisted vulnerability research is moving from theory into practice. For security teams, that raises the bar on how fast they triage findings and patch widely used software. For developers, it is a reminder that legacy code and older dependencies can hide serious risks.

The reported impact is significant, since the flaws could affect 70% of websites. The practical question for the industry is how quickly fixes reach production servers, and whether AI tools will be used equally by defenders and attackers.

What part of your security workflow is most exposed to this shift? Share your view in the comments.