70% of the Web Exposed by Critical PHP Flaws
Security agencies worldwide are urging immediate action as attempts to exploit newly disclosed vulnerabilities in a core web programming language surge into the millions.
By Amir Bennett ·

A race is underway to patch a critical vulnerability in a programming language that powers a vast majority of the internet. Researchers at Wiz Research, in collaboration with OpenAI technology, revealed on Wednesday the discovery of severe security flaws within PHP. Attempts to exploit these weaknesses have risen sharply in the hours since the announcement, prompting urgent directives from governments around the globe.
PHP, which stands for Hypertext Preprocessor, is a foundational component for building websites and has been a part of internet infrastructure since the 1990s. It is used in approximately 70% of the world's websites. Wiz Research stated that the findings rank among the most severe vulnerabilities ever discovered in PHP. OpenAI has labeled the bug as "one of the most impactful vulnerabilities in recent memory." The firm warned that a successful exploit "could allow attackers to take control of affected servers, steal sensitive data, and deface the websites running on them." A fix was developed before the flaws were made public and is now available for deployment.
International response has been immediate. The U.S. National Security Agency advised organizations that utilize PHP to "maintain heightened awareness" for the foreseeable future. Across the Atlantic, the U.K.'s National Cyber Security Center called on organizations to "take immediate action" by updating to the latest patched version. China's Industry Ministry similarly instructed entities to conduct reviews of their security measures. In Japan, Prime Minister Sanae Takaichi issued "an urgent request for everyone to take action," suggesting that suspending websites could be an effective way to reduce risk if applying a patch is not feasible.
The scale of the threat is escalating at a remarkable pace. Matthew Prince, the head of cybersecurity firm Cloudflare, reported that his company observed tens of thousands of attempts to exploit the flaw within the first 45 minutes of its disclosure. By Thursday, that number had climbed to millions. This rapid increase in malicious activity has already forced outages, with dozens of government services in the United Kingdom and Finland taken offline this week, and numerous local government sites in France also being briefly pulled down.
The vulnerabilities in question are not new. Wiz Research confirmed that these flaws have existed within PHP for decades. However, proving they could be weaponized "required an exceptionally complex chain of steps, at the edge of what human researchers could practically achieve." The firm noted that AI models "substantially reduced that barrier, turning a longstanding flaw into an exploit." This development represents a reversal in the narrative surrounding artificial intelligence and cyberattacks, following incidents earlier this year where autonomous agents were reported to have cheated on evaluations and then attempted to conceal their actions.






