The Limits of Differential Privacy (and its Misuse in Data Release and Machine Learning)

2020/11/04 by Josep Domingo-Ferrer, David Sánchez, Domingo-Ferrer, Josep +3 · 4 citations
Computer Science · Engineering · #Age of Information Optimization #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Privacy-Preserving Technologies in Data #Vehicular Ad Hoc Networks (VANETs) #cs.CR

paper · pdf · doi:10.48550/arxiv.2011.02352

Communications of the ACM, to appear

arxiv created 2020/11/04 · openalex publication_date 2020/11/04 · arxiv updated 2020/11/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Differential privacy (DP) is a neat privacy definition that can co-exist with certain well-defined data uses in the context of interactive queries. However, DP is neither a silver bullet for all privacy problems nor a replacement for all previous privacy models. In fact, extreme care should be exercised when trying to extend its use beyond the setting it was designed for. This paper reviews the limitations of DP and its misuse for individual data collection, individual data release, and machine learning.

Cited by

• Privacy Mechanism Design based on Empirical Distributions
• Generative Data Refinement: Just Ask for Better Data
• Shapley-Value-Based Feature Attribution for Data Masking
• Train Once, Forget Precisely: Anchored Optimization for Efficient Post-Hoc Unlearning

Related

• A Comprehensive Guide to Differential Privacy: From Theory to User Expectations
• Differential Privacy: What is all the noise about?
• Practical Considerations for Differential Privacy
• Differential Privacy in Privacy-Preserving Big Data and Learning: Challenge and Opportunity
• Differential Privacy Overview and Fundamental Techniques