The gap between AI adoption and AI security is widening, leaving even the most mature security programs vulnerable to novel exploit chains. While most organizations are rushing to integrate LLMs, few have recalibrated their continuous offensive security testing strategies to account for how these tools change the attack surface. Join Gunter Ollmann, CTO at Cobalt, as he dissects the latest research into how the world's leading security teams are—and aren't—adapting their pentesting cycles for the era of machine-speed threats.