Course Resources

Quick reference materials for the AI Security course. Use these to look up a term, trace a claim back to its source, and check whether something you read here is still current.

  • Glossary – Key terms defined and cross-linked, organized A-Z. Each entry carries back-links to the sections that introduce the term, attack it, and defend it, so an entry is also a map of where a concept lives in the course.
  • References – Every standard, framework, whitepaper, research paper, CVE and tool the course cites, grouped by type, with a note on each explaining what it actually says and where the course uses it.
Read the references list as a currency check, not a bibliography

This field moves faster than the course can. The OWASP LLM Top 10 renumbered eight of its ten identifiers in August 2026; the EU AI Act’s high-risk deadlines moved in July 2026; MITRE ATLAS ships monthly and has added roughly a hundred techniques in a year. Every entry in the references list is version-pinned and dated for that reason.

Two habits the course asks you to keep: cite identifiers with their version (v1.0-C9.2.5, not C9.2.5), and check a framework’s own source rather than a write-up about it – several widely circulated summaries of both AISVS and the Agentic Top 10 state counts and category names that do not match the published standards.