Two-Year-Old PHP Flaw Turns Critical as AI Researchers Build Exploits in 14 Days
A long-standing memory management error in the language powering 70% of the web was turned into a remote attack vector by AI-assisted auditors, forcing an emergency patch release.
By Layla Fontaine ·
The internet’s most ubiquitous open-source language, PHP, is facing an urgent security reckoning after researchers demonstrated that a flaw dormant for two years could be weaponized into a critical remote exploit within just two weeks. While human developers had struggled for months to find a fix for the underlying issue, AI-assisted teams moved with speed to turn theoretical weaknesses into practical attacks, raising questions about the changing pace of cybersecurity defense.
OpenAI announced on Wednesday that researchers under its “Patch the Planet” initiative used its GPT-5.6 Sol model to identify two critical vulnerabilities in PHP, the programming language that powers an estimated 70% of websites. The discovery triggered a rapid, coordinated response involving PHP maintainers, security firms, and government agencies. Although the patch release was originally scheduled for today, Oct. 22, it was moved up to yesterday, Oct. 21, after premature disclosures on Oct. 20 sparked public concern and urgency among site administrators.
The vulnerabilities, tracked as CVE-2026-109618 and CVE-2026-109619, stem from flaws in how PHP manages memory and handles errors. The first occurs when the engine runs out of allocated memory during a request, leaving behind incomplete data that the system may later mistake for valid instructions. The second involves a fatal error that causes PHP to point to memory that has already been freed, allowing subsequent requests to access or corrupt that space. Both issues carry a critical severity score of 9.2 out of 10, meaning they are highly exploitable and impactful.
The timeline of the discovery highlights a significant shift in security dynamics. The PHP Foundation had known about the underlying flaw since 2024 but could not identify a method to exploit it or a reliable fix. Auditing began on July 30 at the Foundation’s request, with researchers from Calif, a cybersecurity firm and founding partner of the initiative, using OpenAI tools. On Aug. 11, Calif researcher Ryan A. identified the exploitable vulnerability, with OpenAI researchers confirming the finding the same day. Shortly after, another firm, Wiz, independently identified an additional related vulnerability.
The implications are broad because the flaws reside in PHP’s core engine rather than in specific application code. This means websites can be vulnerable even if their own code is secure. Attackers can exploit these weaknesses remotely without login credentials to execute malicious code, potentially taking control of servers. The researchers also demonstrated that the flaws could expose sensitive data stored in server memory, including passwords, authentication tokens, and API keys.
OpenAI is framing this incident as a prime example of the “Defender's Window,” a concept the company introduced after its models hacked Hugging Face. The theory posits that there is a narrow timeframe where organizations can use advanced AI to find and patch defenses before those same capabilities become widely available to malicious actors. OpenAI argues that since powerful technology will inevitably be used by bad actors, it is a collective responsibility to act proactively. This stance challenges the counter-argument that flaws which require AI to exploit are effectively harmless until AI becomes standard for attacks.
The PHP Foundation has released patched versions 8.2.35, 8.3.36, 8.4.27, and 8.5.12 to address the issues. Versions 8.1 and older will not receive official fixes, leaving those installations potentially exposed. OpenAI stated that the lengthy two-month fix process was due to the complexity of the task, which required major changes to the PHP engine. The company describes the incident as one of the biggest cybersecurity discoveries of 2026, given the widespread reliance on PHP.
For now, the immediate next step for website owners is to upgrade to the latest patched versions to ensure safety. As AI-driven auditing becomes more common, the gap between discovery and exploitation may continue to shrink, placing increasing pressure on maintainers of foundational software to adapt to this accelerated threat landscape.
