Research
AI uncovers critical PHP vulnerabilities in 70% of web
The Rundown AIhigh confidence
Why it matters
- →AI models are now capable of finding critical vulnerabilities in legacy infrastructure like PHP.
- →The scale of impact is massive, affecting ~70% of websites and prompting government-level responses.
- →This marks a shift from reactive security to proactive AI-assisted code auditing.
OpenAI’s GPT-5.6 Sol model assisted researchers in identifying two critical vulnerabilities in PHP, affecting approximately 70% of websites globally. The National Security Agency has advised organizations to patch immediately or take systems offline to prevent server hijacking. Security firms Calif and Wiz independently verified the flaws using AI models, prompting OpenAI to collaborate with PHP maintainers on patches over two months. This event underscores the growing role of AI in identifying legacy code weaknesses that human auditors have overlooked for years.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
