AI is Uncovering Real Vulnerabilities at Scale

AI Engineer · youtube · 2026-07-13

This talk addresses a growing reality: **AI models can now discover severe vulnerabilities in production code with minimal human effort**. These aren't toy examples, but heavily audited, real-world libraries. Speaker Jack Cable will explain from a security team's perspective: - Where the current attack surface lies - Which existing defense mechanisms work and which fall short - How organizations already using AI coding tools can quickly reduce their risk radius - Why continued improvements in open-weight models will turn this into an issue "everyone must face" Overall, this is a practical AI security discussion rather than mere fear-mongering.

Original post →

More from Safety

Safety channel →