Evaluation & Governance

Membership Inference

An attack that determines whether a specific record was in the training set of a model. Privacy concern for models trained on personal data. Mitigations: differential privacy training, careful eval on member/non-member disparity.

Related terms

Model Extraction
An attack where a black-box model's capabilities are approximated by an adversary through systematic querying. Concern for proprietary hosted models. Rate limiting, query-pattern anomaly detection, and watermarking are common defenses.

Next steps

Take the assessment— See where you stand

Where this fits

Membership Inference is part of the Evaluation & Governance vocabulary used in the Generative AI Maturity Framework. See the full glossary for the complete set of 149 defined terms, or take the free maturity assessment to see where your organisation stands.

Related terms

Related on this site

Framework dimensions

Free tools

Whitepapers

Comparisons

Next steps

Where this fits