← Back to ATLAS MatrixTactic 7 of 15

Execution

The adversary is trying to run malicious code on your AI system.

4 Techniques
MITRE ATLAS Framework

Attack Techniques

Explore the specific techniques adversaries use during the execution phase of AI attacks.

1
AML.T0011

User Execution

Adversaries may rely on user execution to run malicious code in AI environments.

Examples:

  • •Executing malicious scripts in AI environments
  • •Running unauthorized code in ML pipelines
  • •Abusing AI system APIs

Mitigations:

  • ✓User training and awareness
  • ✓Execution controls
  • ✓Monitoring and alerting

References:

MITRE ATLAS AML.T0011
2
AML.T0050

User Execution

Adversaries may rely on user execution to run malicious code in AI environments.

Examples:

  • •Executing malicious scripts in AI environments
  • •Running unauthorized code in ML pipelines
  • •Abusing AI system APIs

Mitigations:

  • ✓User training and awareness
  • ✓Execution controls
  • ✓Monitoring and alerting

References:

MITRE ATLAS AML.T0050
3
AML.T0051

User Execution

Adversaries may rely on user execution to run malicious code in AI environments.

Examples:

  • •Executing malicious scripts in AI environments
  • •Running unauthorized code in ML pipelines
  • •Abusing AI system APIs

Mitigations:

  • ✓User training and awareness
  • ✓Execution controls
  • ✓Monitoring and alerting

References:

MITRE ATLAS AML.T0051
4
AML.T0053

User Execution

Adversaries may rely on user execution to run malicious code in AI environments.

Examples:

  • •Executing malicious scripts in AI environments
  • •Running unauthorized code in ML pipelines
  • •Abusing AI system APIs

Mitigations:

  • ✓User training and awareness
  • ✓Execution controls
  • ✓Monitoring and alerting

References:

MITRE ATLAS AML.T0053
← AI Attack StagingATLAS MatrixPersistence →