Framework guide

MITRE ATLAS for AI Defenders

MITRE ATLAS (Adversarial Threat Landscape for AI Systems) is a knowledge base of tactics, techniques, and case studies for adversarial machine learning. Modeled on MITRE ATT&CK. Used by defenders to model AI-specific threats.

Publisher
MITRE
Kind
Knowledge base of adversary tactics and techniques
Modeled on
MITRE ATT&CK
In the check
Cited by 2 of the 16 questions
01
Reconnaissance
02
Resource Development
03
Initial Access
04
ML Model Access
05
Execution
06
Persistence
07
Privilege Escalation
08
Defense Evasion
09
Credential Access
10
Discovery
11
Collection
12
ML Attack Staging
13
Exfiltration
14
Impact
Figure 1. The MITRE ATLAS tactics in order, from reconnaissance to impact.
Tactics

The tactics, in order

ATLAS follows the ATT&CK shape: an adversary moves left to right, from learning about your model to doing damage with it.

OrderTactic
01Reconnaissance
02Resource Development
03Initial Access
04ML Model Access
05Execution
06Persistence
07Privilege Escalation
08Defense Evasion
09Credential Access
10Discovery
11Collection
12ML Attack Staging
13Exfiltration
14Impact
Use case

Where teams use it

Threat-modeling sessions, adversarial-test planning, SOC detection content authoring, IR playbook design, board-level threat communication.

Posture Check

Where the check cites it

The AI Posture Check cites MITRE ATLAS among its reference frameworks.

Question the check may askDimensionCitation
Do you red-team AI deployments before production, with a documented method? Prompt OWASP LLM01, MITRE ATLAS
For models you build or fine-tune, are there controls against theft and extraction? Model OWASP LLM04, MITRE ATLAS
Ready when you are

Score yourself against this framework.

Five questions, each citing its source. You get your stage, your place on the chart and the one move that matters next.

  • A few minutes for most people
  • Free, from CWS
  • Your stage, the chart and the next move, by email or live with an engineer