A.I. Detects Security Flaw That Could Be Road Map for Hackers to Disrupt the Web

By the Editorial Team · Thursday, October 22, 2026

Security researchers, using advanced AI tools, identified critical vulnerabilities in PHP that could enable remote code execution and widespread website compromises. The urgent release of security patches and global precautionary measures highlight the escalating cybersecurity risks posed by AI-accelerated threat discovery, emphasizing the need for immediate updates and proactive defenses.

Security researchers have uncovered critical vulnerabilities in PHP, a widely used open-source programming language that runs approximately 70 percent of the world’s websites and web applications. The flaws, discovered with the assistance of artificial intelligence tools including OpenAI’s GPT-5.6 Sol and Google’s Gemini 4 models, could allow attackers to remotely execute malicious code on unpatched servers, potentially taking control of websites, stealing data, or disrupting services.

The revelation of these vulnerabilities, dubbed “php2shell” by some experts, prompted an urgent global response on Wednesday, with multiple government agencies, universities, local authorities, and businesses taking websites offline as a precautionary measure. Among those affected were councils in the United Kingdom such as Edinburgh, Oxford, and Bristol, as well as government sites in countries including France, Finland, the Netherlands, and Canada. Innovation, Science and Economic Development Canada temporarily took down its website for urgent maintenance, while European governments also reported service interruptions.

The nonprofit PHP Foundation, which maintains the language, released security patches on Tuesday and urged all users to install updates immediately. However, officials warned that the very public release of the patches could be exploited by hackers to reverse engineer the fixes and develop offensive tools before organizations have the chance to protect their systems. The U.S. National Security Agency (NSA) recommended that websites unable to apply updates promptly consider temporarily disconnecting vulnerable servers to mitigate the risk of attacks.

Experts emphasized the unique challenge posed by AI-related technologies, which accelerate both the identification of software flaws and the development of attack methods. “Given the severity of these flaws and PHP’s widespread use, we consider this one of the most impactful security issues in recent memory,” said OpenAI, underscoring the dual-use nature of AI in cybersecurity.

Cybersecurity leaders, including Rajiv Gupta, head of the Canadian Centre for Cyber Security, indicated that the discovery marks a growing trend of vulnerabilities emerging as AI capabilities advance. Gupta advised organizations to monitor logs for suspicious activities and recommended temporarily taking vulnerable web servers offline to prevent exploitation.

Security firms noted that although patches are available, a significant portion of PHP servers remain unpatched, creating a large attack surface that could lead to ongoing risk for years. “Seventy to 80 percent of the web runs on PHP. The wave hasn’t even hit us yet,” said David Shipley, CEO of Beauceron Security Inc.

The discovery follows increased attention to AI-driven cybersecurity risks, as both defenders and attackers harness artificial intelligence to identify and exploit system weaknesses. The PHP vulnerabilities have intensified calls for heightened vigilance and proactive patch management across the global internet infrastructure to prevent widescale disruption.

This article was reported by synthesizing coverage from multiple independent news outlets to present a balanced, unbiased account. Read about our editorial standards. Spotted an error? Let us know.