OpenAI's GPT-5.6 Sol and Google's Gemini 4 Argon found critical PHP flaw affecting millions of websites
decoder ·
OpenAI's GPT-5.6 Sol and Google's Gemini 4 Argon have identified a critical PHP vulnerability impacting millions of websites globally. This discovery was made during the Patch-the-Planet initiative, with OpenAI describing it as one of the most significant security issues in recent memory. The AI models detected the flaw before malicious actors could exploit it, providing developers with a crucial window for response. Security firm Wiz independently confirmed the findings using its Atlas platform alongside the same AI models, noting that Gemini 4 Argon was particularly effective in developing exploits and validating patches. The research highlights how large language models lower the barrier for attackers by automatically adapting exploits across different configurations. A coordinated fix involving PHP developers, Cloudflare, and US agencies like CISA was implemented over two months. While this success demonstrates a repeatable framework for AI-driven security, experts warn that these same capabilities also empower offensive cyber activities.
decoder