Critical PHP bugs expose web's fragility as AI speeds up exploitation
· Tech
Two critical vulnerabilities in the PHP web programming language, which underpins a large share of the internet, threatened to disrupt services across the web, according to TechCrunch and WIRED. Security researchers used AI tools to help uncover the years-old flaws, and AI-assisted attacks began shortly afterward, the reports said. The researcher who set off the chain reaction said it put roughly 70% of the web at risk. Both reports frame the episode as unlikely to be the last major software scare, and highlight how AI is making existing bugs faster and easier to exploit. The incident matters because PHP remains one of the most widely deployed languages for websites and web applications. The speed with which flaws can now be found and weaponized compresses the window organizations have to patch. The coverage points to a broader shift in software security, where AI tools cut both ways: helping defenders find hidden bugs, but also enabling attackers to move faster once a flaw is public.
Why it matters:The episode shows how AI is compressing the time between vulnerability discovery and exploitation, raising the pressure on businesses that rely on widely used open-source software to patch faster.
Key points
- Two critical bugs in PHP, a language used by most of the internet, triggered the scare.
- Researchers used AI tools to uncover years-old vulnerabilities before AI-assisted attacks began.
- The researcher behind the disclosure said about 70% of the web was put at risk.
Sources (original reporting)
More stories from this day
- Starlink service in Ukraine restored after 10-day throttling
- Pennsylvania law requires labels on AI-generated political ads
- OpenAI undercuts Anthropic on price, winning back coding work
- Study finds Grok and ChatGPT recommend opposing political candidates
- Credit markets spooked by 'total meltdown' in riskiest debt; metals traders post record profits
- Bannerman-CNNC joint venture takes FID on Namibia uranium project