Create Proxy AI Model

Description

Adversaries may obtain models to serve as proxies for the target model in use at the victim organization. Proxy models are used to simulate complete access to the target model in a fully offline manner. Adversaries may train models from representative datasets, attempt to replicate models from victim inference APIs, or use available pre-trained models.

Honesty-tier rationale

This step needs real training/compute (e.g. training a proxy model); only the outcome can be visualized in-browser.

Sub-techniques

• AML.T0005.000 — Train Proxy via Gathered AI Artifacts
• AML.T0005.001 — Train Proxy via Replication
• AML.T0005.002 — Use Pre-Trained Model

Source

• Official MITRE ATLAS page: AML.T0005