From Statistical Disclosure Control to Fair AI: Navigating Fundamental Tradeoffs in Differential Privacy

TL;DR

This paper explores the fundamental tradeoffs between privacy, utility, and fairness in differential privacy, providing a unified framework to understand limits and guide practical decision-making in fair AI systems.

Contribution

It systematically connects privacy, fairness, and utility tradeoffs, characterizing the Pareto frontier and offering practical guidance for private fair machine learning.

Findings

Demonstrates the Pareto frontier between privacy, utility, and fairness.

Characterizes fundamental limits on achieving all three simultaneously.

Provides practical guidance for deploying private fair learning systems.

Abstract

Differential privacy has become the gold standard for privacy-preserving machine learning systems. Unfortunately, subsequent work has primarily fixated on the privacy-utility tradeoff, leaving the subject of fairness constraints undervalued and under-researched. This paper provides a systematic treatment connecting three threads: (1) Dalenius's impossibility results for semantic privacy, (2) Dwork's differential privacy as an achievable alternative, and (3) emerging impossibility results from the addition of a fairness requirement. Through concrete examples and technical analysis, the three-way Pareto frontier between privacy, utility, and fairness is demonstrated to showcase the fundamental limits on what can be simultaneously achieved. In this work, these limits are characterized, the impact on minority groups is demonstrated, and practical guidance for navigating these tradeoffs are provided. This forms a unified framework synthesizing scattered results to help practitioners and policymakers make informed decisions when deploying private fair learning systems.