AIRiskAware
AI Governance Glossary
Governance Practice

What Is Differential Privacy?

Differential Privacy is a mathematical privacy guarantee that limits the amount of information that can be inferred about any individual from an AI model or dataset, by adding calibrated statistical noise.

Definition

Differential Privacya mathematical privacy guarantee that limits the amount of information that can be inferred about any individual from an AI model or dataset, by adding calibrated statistical noise.

Differential privacy (DP) is a rigorous formal definition of privacy that can be proven mathematically. A DP mechanism guarantees that the probability of any output changes by at most a factor of e^ε when any single individual's data is added or removed — ε (epsilon) is the privacy budget. Lower epsilon means stronger privacy but typically worse model utility. DP is used in practice by Google (Chrome RAPPOR), Apple (iOS telemetry), Meta (ad measurement), and the US Census Bureau. It is increasingly referenced as a technical control in AI governance frameworks.

Source: Dwork & Roth (2014); NIST SP 800-226

Plain-language explanation

Differential privacy (DP) is a rigorous formal definition of privacy that can be proven mathematically. A DP mechanism guarantees that the probability of any output changes by at most a factor of e^ε when any single individual's data is added or removed — ε (epsilon) is the privacy budget. Lower epsilon means stronger privacy but typically worse model utility. DP is used in practice by Google (Chrome RAPPOR), Apple (iOS telemetry), Meta (ad measurement), and the US Census Bureau. It is increasingly referenced as a technical control in AI governance frameworks.

Primary source: Dwork & Roth (2014); NIST SP 800-226

See where you stand on AI governance

Take the free 7-question maturity assessment and get a personalised action plan.

Free assessment — 3 minutes →