Markets and Incentives for Transacting on (Personal) Information, Part III

on YouTube ↗01h 05m

" summary

Differential privacy provides a formal guarantee that an epsilon change barely changes individual expected utility, so mechanisms satisfying epsilon-differential privacy => approximate truthfulness make truthful reporting nearly a dominant strategy and enable new mechanism-design constructions (digital goods auctions, equilibrium selection, joint-DP outputs). Pricing personal data proves thorny because payment needs and harms are often cost correlated with content, opening impossibility results, arbitrage and adverse-selection risks. Lab-style experiments generate synthetic sensitive signals to study behavior, and concerns about societal impacts and composability (census use case) remain.

AI-written summary · about

" chapters (23)

1. 0:00Intro
2. 1:16(Private) data as a good
3. 1:51Data vs. Sneakers
4. 9:56Privacy: What is the good?
5. 13:30What should a data price cover?
6. 15:23Supply and demand
7. 16:08Talk Outline
8. 16:38Revealed preferences
9. 22:20"Privacy and Mechanism Design" Roth and Pai SigEcomExchanges 2013
10. 27:59DP as a tool for mechanism design
11. 30:37DP for mechanism design
12. 32:27Eliciting private data: Buying Private Data With Verification
13. 33:43Responding to impossibility
14. 36:53Measuring the costs of privacy
15. 39:34Impacts of privacy concerns
16. 41:03Design of a data market
17. 41:33Two econ lessons on how to design experiments
18. 50:20The privacy regulation hurdle
19. 51:19Philosophical/legal/moral benefits of privacy
20. 52:58Privacy protections prevent harms
21. 53:47More data, better data, more access
22. 54:36Privacy = Statistical validity
23. 55:55Other lessons from economics

" related talks

• 19mPractical Considerations for Differential PrivacyAlex Kulesza · PEPR '25
• 30mAccuracy First: Selecting a Differential Privacy Level for Accuracy-Constrained...Katrina Ligett · Optimization, Statistics and Uncertainty
• 06mLessons Learned in Communicating Differential Privacy to Data SubjectsMary Anne Smart · SOUPS 2023
• 20mDesigning for User Privacy: Integrating Differential Privacy into Ad Measurement...Shikun Zhang, Jian Du · PEPR '24
• 42mDifferential PrivacyKunal Talwar · Uncertainty in Computation 2016
• 01h 00mPANEL: Statistical Theory, Privacy and Data AnalysisThomas Steinke, John Duchi, Anand Sarwate, Yu-Xiang Wang, Aleksandra Slavkovic, Bin Yu · Privacy and the Science of Data Analysis 2019
• 17mNegotiating Privacy/Utility Trade-Offs under Differential PrivacyGerome Miklau · PEPR '22
• 01h 00mDifferential Privacy: Fundamentals to Forefront ICynthia Dwork · Cryptography Boot Camp 2015