PHP Flaws Found via AI, Fixed in 24 Hours by Single Maintainer

Two critical security vulnerabilities in PHP, used by ~70% of websites, were discovered by researchers using an OpenAI AI model and patched within 24 hours by maintainer Ondřej Surý. The incident highlights the growing role of AI in vulnerability discovery and the fragility of open-source maintenance, where a single individual bears critical security responsibility. The full scope of the breaches and any exploitation remain undisclosed.

AI disclosure: Model-written summary of the source, not a verified fact. Check the sources. How Mosaic works