OpenAI Reveals Critical PHP Flaws Exploited via AI in Two Weeks

The tension between rapid AI-driven discovery and the slower pace of legacy software maintenance reached a breaking point this week when OpenAI disclosed that its AI model, GPT-5.6 Sol, had identified critical security vulnerabilities in PHP. The programming language, which underpins an estimated 70% of the internet, contained two severe flaws that had persisted for two years despite human scrutiny. The announcement highlights a growing divide in cybersecurity: while human developers struggled to resolve the underlying code issues, AI-assisted researchers converted the dormant bugs into working exploits in a matter of weeks.

The path to this disclosure began in 2024, when the PHP Foundation first became aware of the underlying defect. For over a year, maintainers and security teams examined the code but failed to identify a method for exploitation or a reliable fix. It was not until July 30, 2026, that the Foundation requested an audit using OpenAI’s tools as part of the 'Patch the Planet' cybersecurity initiative. The audit was led by Calif, a California-based cybersecurity firm and founding partner of the program.

On August 11, 2026, the audit yielded its first results. Calif researcher Ryan A. identified an exploitable vulnerability, and OpenAI researchers confirmed the finding the same day. Shortly thereafter, another cybersecurity firm, Wiz, independently discovered a second, related vulnerability. These two issues, designated CVE-2026-109618 and CVE-2026-109619, were traced to fundamental errors in how PHP manages memory and handles processing errors.

The technical nature of the flaws allows for remote exploitation without login credentials. The first vulnerability occurs when PHP exhausts its allocated memory during a request, potentially leaving incomplete internal data that the engine later misinterprets as valid. The second involves a fatal error that leaves the system pointing to memory that has already been freed, allowing subsequent requests to access or corrupt that space. Under specific conditions, attackers can leverage these weaknesses to execute malicious code on affected servers, potentially gaining control of websites and exposing sensitive data such as passwords, authentication tokens, and API keys.

Both vulnerabilities carry a critical severity score of 9.2 out of 10. Because the flaws reside in PHP’s core engine, any website using the language is potentially affected, regardless of the security quality of its specific application code. OpenAI noted that while human teams could not fully resolve the issues, AI-wielding researchers demonstrated practical exploitation methods within two weeks of identification.

In response to the findings, a coalition of researchers, PHP maintainers, security providers, and government cybersecurity agencies coordinated a two-month effort to develop and deploy fixes. The PHP Foundation released patched versions 8.2.35, 8.3.36, 8.4.27, and 8.5.12 to address the vulnerabilities. Versions 8.1 and older will not receive official fixes, meaning those installations require immediate upgrades to remain secure.

The coordinated release was originally planned for October 22, 2026. However, following premature disclosures on October 20 that triggered public concern, the release was brought forward to October 21. OpenAI stated that the extended timeline was necessary due to the complexity of the task, which required significant changes to the PHP engine to ensure a stable and secure resolution.

OpenAI framed this incident as a prime example of its 'Defender's Window' concept. The company argues that there is a narrow timeframe during which organizations can leverage advanced frontier AI to find and patch defenses before those same capabilities become widely available to malicious actors. While some argue that flaws requiring AI to exploit are effectively harmless, OpenAI maintains that proactive defense is a collective responsibility, especially as competitors like Anthropic’s Claude Mythos have recently identified thousands of zero-day vulnerabilities across major operating systems and browsers.

The situation underscores the shifting landscape of software security. As AI models become more capable at auditing complex codebases, the window for defenders to act is shrinking. For the Mid-Atlantic and global web communities, the immediate priority is clear: verify PHP versions and upgrade to the latest patched releases to mitigate the risk of remote exploitation.