With the emergence of autonomous AI agents – also known as "Agentic AI" - new threat landscapes are introduced. This talk will explore a new approach on how to tackle the cybersecurity challenges: combining the MAESTRO framework, a newly published Agentic AI threat modeling framework, with MITRE's D3FEND. The goal is to build stronger, more adaptable defenses against these cutting-edge threats.

Here are the key discussion points:

An Introduction and Deep Dive into MAESTRO: We'll introduce the MAESTRO framework. MAESTRO provides a structured, multi-layered approach for spotting and mitigating risks. It is all about being proactive in securing agentic AI.

The Unique Threat of Agentic AI: We'll kick things off by exploring the specific security risks that come with the Agentic AI systems. We will discuss the reason why simply relying on our go-to threat modeling approaches isn't enough anymore.

Understanding MITRE D3FEND: Next, we'll provide a practical overview of the D3FEND framework. We will explore its catalog of defensive cybersecurity techniques and its standardized terminology for describing and countering cyber threats. We will be explaining D3FEND in the context of developing solid defense mechanisms.

Putting It All Together: Integrative Strategies: The core of the presentation! We'll demonstrate how to connect the dots between threats identified using MAESTRO and the specific defensive techniques outlined in D3FEND. The goal is to propose a unified, comprehensive security strategy specifically tailored for Agentic AI systems.

Attendees will leave with a clear understanding of how to combine cutting-edge agentic AI threat modeling with proven defensive frameworks. This knowledge will help the audience to significantly boost the confidence in the context of cybersecurity for their Agentic AI systems deployment.