S4E10 Resources: The Model Is the Target
Resources from Season 4 Episode 10, “The Model Is the Target”. AI changes what organisations need to protect, how attacks can arrive, and what security testing has to prove.
This page covers changed AI crown jewels, the AI operating context, direct and indirect prompt injection, the Linwei Ding, EchoLeak, and OpenAI and Hugging Face case studies, AI assurance, red teaming, agent permissions, containment, logs, review points, kill switches, a practical three-question checklist, and a plain-English glossary.