Xataka warns that AI is accelerating the aging of internet software
🎧 Listen to this story
Xataka warns that the infrastructure supporting the internet is aging and that the arrival of frontier artificial intelligence models may be accelerating that process. In PHP’s case, the PHP Foundation identified a vulnerability in 2024 but found no known way to exploit it at the time; the difficulty of fixing it also delayed a solution, the site says. The problem is worsened by dependence on old applications, libraries and operating systems, whose updates can break services that are still running.
According to Xataka, AI can help review large volumes of code and find patterns that are difficult to detect through conventional methods. Google Cloud described how its engineers used AI models to locate and validate PHP flaws, create exploitation tests and verify fixes — although identifying vulnerabilities does not remove the difficulty of applying them across a broad and heterogeneous infrastructure.
Why it matters · editorial interpretation
The case suggests that finding vulnerabilities with AI is only one step: risk remains when fixing old systems can break services and requires validation across heterogeneous infrastructure. For development teams, this reinforces the need to combine automated review with testing and careful upgrade planning.
Sources
SecurityCode