Explore a conference talk from Security BSides London that delves into leveraging Large Language Models for dynamic threat modeling in cybersecurity. Learn how traditional static architecture diagram-based threat modeling falls short as applications evolve, and discover innovative approaches to maintaining current threat models. Understand the challenges security teams face in keeping threat assessments up-to-date and examine potential solutions using AI technology to create more responsive and adaptive security frameworks. The 18-minute presentation demonstrates how organizations can better protect themselves against emerging threats by moving beyond conventional static modeling methods to more dynamic, AI-driven approaches.