OpenAI and Wiz AI Models Uncover Critical Global PHP Security Flaws

22 Oct 2026, 2:59 am · 9h ago · 1 min read · Inc42

Critical security vulnerabilities discovered in the PHP programming language have triggered emergency advisories from cybersecurity agencies worldwide, including India's CERT-In and the US NSA. The flaws, which allow attackers direct access to server file systems and databases, were independently flagged by artificial intelligence systems, specifically OpenAI's GPT-5.6 Sol model and Google-owned cybersecurity firm Wiz under the "Patch the Planet" initiative. Multiple governments and platforms, including the Wikimedia Foundation, took systems offline or patched servers rapidly. The discovery marks a major milestone for AI agents autonomously identifying complex, real-world zero-day vulnerabilities across core internet infrastructure.