In brief
A breakthrough by an artificial‑intelligence system revealed a dangerous weakness in PHP, the engine behind a large share of the internet. The flaw rattled governments and cyberspace specialists as they rushed to patch servers that power everything from online stores to encyclopedias. The cyber‑security group Wiz labeled it “the most severe ever found,” while OpenAI’s creators said it was among the most impactful discoveries in recent memory. Experts warn that the incident underscores the increasing role of AI in both hunting and potentially exploiting vulnerabilities.
Artificial‑intelligence has uncovered a glaring flaw that endangers more than 70% of the world’s websites, the share of sites run on PHP. The discovery was made by a model trained to scan code for hidden risks, and it flagged issues that could let attackers hijack websites or siphon private data.
PHP powers high‑traffic sites such as Wikipedia, major e‑commerce platforms, and numerous government portals. The identified weaknesses had been present in the software for years, only now exposed by AI‑driven analysis. In response, the PHP development community released a rapid patch that addressed the vulnerabilities and urged site owners to update their installations immediately.
Cyber‑security firm Wiz has described the flaw as the “most severe” discovered to date, noting that its exploitation could compromise critical digital infrastructure. OpenAI, the organization whose tools were used in the search, called the finding “one of the most impactful security issues in recent memory.”
Governments and security agencies across the globe are now scrambling to assess the extent of the risk, evaluate whether existing defenses can contain potential breaches, and coordinate response strategies with web operators.
For users, the main takeaway is to watch for updates from site administrators, ensure that personal data is not stored in unsecured systems, and stay informed through official security advisories.
