Lessons Learned in Communicating Differential Privacy to Data Subjects
summary
Two empirical studies evaluate how explanations about differential privacy affect understanding and sharing decisions, comparing local vs central models (noise added on-device versus at the server) and testing explanations about the epsilon tradeoff. Key findings: concrete statements about what privacy means improve comprehension; people weigh privacy and utility and want utility consequences reported; a clear explanation can reduce willingness to share; and risk comprehension is a better evaluation metric than simple share rates.
AI-written summary · about
related talks
10mCommunicating Differential Privacy Guarantees to Data Subjects
01h 05mMarkets and Incentives for Transacting on (Personal) Information, Part III
09mReplication: The Effect of Differential Privacy Communication on German Users'...
01h 00mPANEL: Statistical Theory, Privacy and Data Analysis
11mExploring User-Suitable Metaphors for Differentially Private Data Analyses
17mNegotiating Privacy/Utility Trade-Offs under Differential Privacy
19mPractical Considerations for Differential Privacy
30mAccuracy First: Selecting a Differential Privacy Level for Accuracy-Constrained...