MITRE ATLAStechnique

AML.T0057 LLM Data Leakage

Published 2023-10-25 Modified 2023-10-25 Official page ↗
Tactics
Exfiltration
Created
2023-10-25

Description

Adversaries may craft prompts that induce the LLM to leak sensitive information. This can include private user data or proprietary information. The leaked information may come from proprietary training data, data sources the LLM is connected to, or information from other users of the LLM.

Related

Tactic · MITRE ATLAS (1)

Referenced by MITRE ATLAS · mitigates (4)

  • AML.M0020 Generative AI Guardrails mitigation
  • AML.M0021 Generative AI Guidelines mitigation
  • AML.M0022 Generative AI Model Alignment mitigation
  • AML.M0008 Validate AI Model mitigation