Speaker: Gerd Giese

Artificial intelligence (AI) is transforming not only defense, but increasingly the offensive side as well. In this 15-minute live demo, we show in a controlled lab environment how freely available or affordable Large Language Models (LLMs) can already be used today to support and partially automate key steps in a cyberattack.

Using a deliberately vulnerable target system, we demonstrate how an agentic AI approach gathers information, evaluates results, derives next steps, and thereby significantly accelerates the attack process. The focus is not on the technical details of individual attack methods, but on the overarching insight: the barrier to entry for more sophisticated cyberattacks is dropping significantly because AI makes capabilities acc ...