Threat intelligence / Technique

Models

Adversaries may publish poisoned AI models through model registries, code repositories, or other model distribution channels. The model may be newly created or a modified variant of a legitimate model and may contain manipulated weights, configurations, architecture, serialized code, or other components that produce malicious behavior or execute malicious code. Victims may subsequently download and integrate the model through [AI Supply Chain Compromise](/techniques/AML.T0010).

Technique

TypeTechnique

External ID—

First seen2026-08-14

Last seen2026-08-14

Confidence

—

Adversaries may publish poisoned AI models through model registries, code repositories, or other model distribution channels. The model may be newly created or a modified variant of a legitimate model and may contain manipulated weights, configurations, architecture, serialized code, or other components that produce malicious behavior or execute malicious code. Victims may subsequently download and integrate the model through [AI Supply Chain Compromise](/techniques/AML.T0010).

8 recorded

Relationships

mitigates

from Validate AI Model

Mitigation

mitigates

from Vulnerability Scanning

Mitigation

subtechnique-of

to Publish Poisoned AI Artifacts

Technique

uses

from Malicious Models on Hugging Face

Campaign

uses

from Model Namespace Reuse Supply Chain Attack

Campaign

uses

from Organization Confusion on Hugging Face

Campaign

uses

from Poisoned GGUF Templates: Inference-Time Supply Chain Attack

Campaign

uses

Campaign

1 external

References

mitre-atlas

AML.T0115.001