The Top Security Vulnerabilities Generated by AI Code

What Al-generated apps get wrong: a vulnerability study across models, vendors, languages, and a real-world app

More engineering teams are shipping AI-generated code faster than security teams can review it. Every industry report says the same thing: AI code ships more bugs. What's missing is specificity: Which bugs, why and what you can do about it?

So we tested it. Using Xint, our autonomous pentesting platform, we scanned source code from apps built three different ways: from a detailed spec, from a one-line "just build me this" prompt, and by hardening a real, mature production codebase. Every finding in this report was individually validated and proven exploitable — not raw scanner noise.

‍