SEC536: Adversarial AI - Penetration Testing AI Systems

In-Person
The world is connecting AI to everything, yet we are still learning how to secure AI systems effectively. This creates new risks for society and puts significant pressure on security professionals. This session is here to help.
Rob will cover the essentials of AI security and offer an exclusive sneak peek into two upcoming AI security standards: ISO/IEC 27090 and the AI Act cybersecurity standard prEN 18282, for which Rob serves as co-editor.
You will learn what these standards cover and how to start understanding AI security threats and protecting AI systems using the OWASP AI Exchange flagship project as a practical resource. The Exchange has provided substantial input to both standards and helps translate emerging international standardization into concrete security guidance. He will also touch on the partnership between the OWASP AI Exchange and SANS Institute, and how this work contributed to the SANS Critical AI Security Guidelines.
By the end of the session, attendees will have a solid understanding of the most relevant AI security threats and actionable guidance for securing AI systems in line with emerging international standards.
In-Person
In a world of agentic AI, how do we make sense of what to use when? In this talk, Jean-François Maes brakes it down, what plugins to use, what mcp to use, what skills, markdowns etc. Give yourself superpowers and get stuff done!
In-Person