Techniqueresource-developmentATLAS

AML.T0058Publish Poisoned Models

What it is

Adversaries may publish a poisoned model to a public location such as a model registry or code repository. The poisoned model may be a novel model or a poisoned variant of an existing open-source model. This model may be introduced to a victim system via [AI Supply Chain Compromise](/techniques/AML.T0010).

References

  1. https://atlas.mitre.org/techniques/AML.T0058

Related by meaning· 6

Nearest entities by semantic similarity across the cs-graph corpus.

ATLAS
Publish Poisoned Datasets
ATLAS
Poison Training Data
ATLAS
Manipulate AI Model
ATLAS
AI Agent Tool Poisoning
ATLAS
AI Agent Tool Data Poisoning
ATLAS
Corrupt AI Model
Sourced from MITRE ATLAS — Adversarial Threat Landscape for AI Systems. Curated by Adam Lundqvist, SQUR.