Threat Modeling AI Systems (225)
from Shostack + Associates
Threat Modeling AI Systems is an in-depth technical course that focuses on how to effectively identify threats and design controls in LLMs and other non-deterministic software. This course is designed for:
This intensive course teaches security professionals how to apply threat modeling to AI systems using a principles-based approach. Rather than focusing on specific AI models or tools that quickly become outdated, the course focuses on the fundamental concepts that drive AI behavior, risk, and security controls. Participants will learn how AI systems work, how attackers exploit them, and why threats such as data poisoning, prompt injection, and model theft require a different security mindset.
What to expect
Using real-world scenarios and collaborative discussions, participants will practice identifying high-impact AI threats, evaluating security controls, and facilitating effective threat modeling conversations across multidisciplinary teams of security engineers, architects, and data scientists.
Prerequisites
Participants should already be familiar with:
For individuals unfamiliar with the Four Question Framework, The World’s Shortest Threat Modeling Course is free, self-paced, and can be completed in under an hour.
Learning outcomes
By the end of the training, participants will be able to:
Timing
This course is currently offered in-person only, over 2 consecutive days (16 hours total instruction time)
Threat Modeling AI Systems is an in-person 2- day course totaling 16 hours total instruction time. The time is split between lectures and group exercises
Skills covered in this course
Topics covered
Threat Modeling AI systems is currently only offered as a 2-day, live instruction course.
At this time, we do not have any upcoming open enrollment course dates scheduled. However, with a minimum enrollment, Threat Modeling AI Systems, can be delivered to your organization as a closed course.