AI Hackers Are Faster Than Your Pen Test — Eli Cohen, Snyk
62% of LLM-generated code is unsafe, while AI accelerates attacker capability simultaneously.
“62% of the results generated by LLM are unsafe or corrupted”
Snyk field CTO Eli Cohen presents data showing AI coding agents have driven a 218% increase in daily code generation, but 62% of LLM-generated outputs contain security flaws. Compounding the risk, attackers are using AI to chain low-severity vulnerabilities into critical exploits, outpacing traditional defensive workflows. The talk frames this as a structural security crisis requiring continuous offensive security practices rather than periodic pen testing.