In brief
Artificial intelligence tools uncovered a set of long‑hidden flaws in the PHP programming language that powers millions of websites, sparking a coordinated response from governments and cyber‑security agencies worldwide. The PHP core team rolled out a patch, but many high‑profile sites were taken offline to protect against unauthenticated attacks that could hijack servers and exfiltrate data. The incident highlights both the power of AI to discover deep‑sea bugs and its potential to accelerate attack development, prompting major tech companies and national authorities to issue emergency alerts and urge operators to patch immediately. Public‑facing sites, including Wikipedia and university portals, have been temporarily shut down or defended, while US, UK, EU, Singapore, Australia and Japan officials have issued guidance to enforce swift defenses.
Artificial intelligence tools uncovered a set of long‑hidden flaws in the PHP programming language, the backbone of millions of websites worldwide, sparking a coordinated response from governments, cyber‑security agencies and major tech companies.
These vulnerabilities, described by security firms Wiz and Calif as the “most severe ever detected by AI”, allow unauthenticated attackers to hijack web servers and exfiltrate database contents.
In response, the PHP core team rushed a patch and urged all users to upgrade immediately. The US NSA and the UK National Cyber Security Centre issued emergency advisories, recommending site outages where patches were not yet installed.
High‑profile sites – including Wikipedia, the University of Geneva, and government portals in Finland and the Netherlands – were temporarily taken offline. The incident has also highlighted the dual nature of AI, capable of both discovering dark corners in code and accelerating attacks.
What This Means for Web Operators
Website owners are urged to apply the latest PHP patch, monitor for suspicious activity and heed any additional guidance from cyber‑security agencies. Users should remain alert to phishing messages exploiting the public‑facing nature of the news.
Global Response
Official alerts have reached every continent: Japan’s prime minister called for checks on public web pages, Singapore, Australia and the EU issued warnings, and Amazon confirmed it had faced multiple intrusion attempts before the disclosure.
