From the course: Advanced AI Threat Modeling and Risk Assessment
Join today to access over 26,000 courses taught by industry experts.
End-to-end view of AI pipeline threat modeling
From the course: Advanced AI Threat Modeling and Risk Assessment
End-to-end view of AI pipeline threat modeling
You've made it to the end of the course. Now, we're going to put all those pieces back together. I'm going to provide a comprehensive end-to-end view of AI pipeline, where the threats lie, we'll revisit some of the key frameworks, and then we'll provide a clear decision logic of when to use each one. So let's start with the end-to-end AI pipeline. It's really where traditional security ends, and we have to look at this in a new way, which is the whole reason why you took this course to start with. It begins with data collection, where the new threats are things like data poisoning, bias injections. Then we talk about model training, where we face risks like backdoor injection and model theft. Some of you may never actually have to deal with this, but you need to be cognizant of the models that you're using could have had this happen as part of that data supply chain issue we talked about previously. Then comes deployment, where we see prompt injections once again and evasion attacks…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.