Security Challenges in Autonomous AI Agents

Learn about the emerging field of autonomous agent security, key attack vectors, design vulnerabilities and more.

As AI systems gain autonomy and tool-using capabilities, their security becomes significantly more complex. This lecture introduces the emerging field of autonomous agent security, highlights key attack vectors and design vulnerabilities, and discusses current research directions for building robust and trustworthy agentic systems.

Benefits for the attendees

• What autonomous AI agents are and how they behave
• The biggest security risks and how attackers can exploit them
•  Real examples of vulnerabilities in agentic systems
•  Practical principles for building and using AI agents safely
•  Where AI security research is heading and why it matters

Who can participate?

The lecture is intended for students, professionals, and members of the public interested in understanding the next major challenge in AI safety.

Prerequisites

Basic interest in AI or modern technology, but no prior knowledge of cybersecurity or programming is required.