Artificial Hackers: Unleashing the dogs of war

This session explores the emergence of AgenticAIs—advanced, autonomous systems capable of executing sophisticated cyberattacks using a network of specialized LLMs and their tooling. Unlike human attackers, Agenics AIs operate without ethical considerations, blowing up the principle of security through obscurity, adapting in real time their TTP and code and making adversary attribution harder. We examine the risks associated with these entities, which are reshaping the landscape of cybersecurity.