Hritesh SonawaneSecurity engineering
NotesAI security engineeringModule 1

Module 1 of 5

The threat landscape

Before controls, a map: what makes AI systems a distinct security problem, how to threat-model one, and which public frameworks to map your work to.

3 lessons41 min reading

Before controls, a map. This module explains what makes AI systems a distinct security problem, gives you a repeatable way to threat-model one, and introduces the public frameworks your work will be measured against.

By the end of it you should be able to point at any AI feature and answer two questions: what can influence the model, and what can the model influence.

Lessons

  1. 1.aWhy AI security is differentWhat is new, and what is ordinary security under a new name?11 min
  2. 1.bThreat modeling an AI systemHow do I find what can go wrong in a specific design?16 min
  3. 1.cFrameworks and mapsWhat do OWASP, MITRE ATLAS, NIST, and the EU AI Act each give me?14 min