In brief
Artificial‑intelligence tools flag a long‑standing vulnerability in PHP, the scripting language used by over 70% of websites, including Wikipedia and major e‑commerce platforms. Security firm Wiz described the flaw as the most severe ever detected, while OpenAI called it one of the most impactful security issues in recent memory. Governments and cyber‑security agencies are rushing to deploy patches, as the vulnerability had lurked in PHP’s codebase for years. The incident highlights the evolving role of AI in uncovering hidden threats and the need for tighter monitoring of core internet infrastructure.
A new artificial‑intelligence audit uncovered a critical flaw in the PHP programming language, the engine powering more than 70% of all sites on the web, from Wikipedia to major online retailers. The vulnerability, which had existed in the code for years, could allow attackers to inject malicious scripts or exfiltrate sensitive data.
Cyber‑security firm Wiz called the problem the most severe ever detected. “This is a zero‑day that has been lurking in production for too long,” said the firm’s chief cybersecurity researcher.
OpenAI, whose GPT‑derived model was used to find the flaw, described the discovery as “one of the most impactful security issues in recent memory.” The firm stated that using generative models to traverse codebases can pinpoint hidden vulnerabilities that traditional static analysis often misses.
Governments and security agencies worldwide are scrambling to deploy emergency patches. The PHP development team has released a patch for version 8.3 and plans updates for older releases by the end of the month.
The incident underscores the growing importance of AI tools in auditing the backbone of the internet and poses new questions about how software developers will keep core technologies secure in an increasingly automated world.
