Adversarial ML & AI-Driven Security: Exploiting and Defending Modern Infrastructure

In this event, we will explore the intersection of Artificial Intelligence and Cybersecurity, analyzing both the vulnerabilities of current models and their use in offensive and defensive operations. The talk will be divided into two main blocks:

• Adversarial ML: Vulnerability analysis in Computer Vision models and LLMs (evasion attacks, jailbreaking...).
• AI-Driven Security: How AI helps detect and prevent threats, and how to use intelligent agents to automate Red Teaming and the defense of critical infrastructure.

We will feature Daniel López Gala, cybersecurity researcher at the Cyber-Defence Campus (armasuisse) and ETH Zürich, who will explain how these concepts are applied in real-world research and defense environments. Event slides