← AI security in plain English

AML.TA0008

MITRE ATLAS / Discovery

The adversary tries to map the AI environment from the inside, using techniques such as Discover AI Model Family, Discover AI Artifacts, and Discover LLM System Information.

Think of it likeOnce inside, wandering room to room to learn the layout: where the valuables sit and which doors lead where.

In plain English

Now inside, attackers explore the AI setup: which model family it is, what data and artifacts exist, and how it is configured, to plan their next move.