The Register says AI vulnerability hunting still cannot beat bad human security habits

The Register reported that AI may be good at finding security vulnerabilities, but it cannot beat human stupidity. The piece is a useful corrective to the idea that Mythos-like cyber models are the whole security story: many real-world compromises still start with poor configuration, weak process, exposed secrets, or users taking unsafe actions. The AI angle matters because better vulnerability discovery can raise patch pressure while leaving the oldest operational risks untouched.

Key details: Published June 29, 2026 by The Register, The article discusses AI vulnerability discovery, It argues basic human and operational failures remain exploitable, The story connects AI security tooling with real-world security practice.

Why it matters: AI security tools may find more bugs, but organizations still lose when basic process and configuration discipline fail.