On the Differential Privacy and Interactivity of Privacy Sandbox Reports

2024-12-22

arXiv.org

AI Summary

This work formally analyzes the privacy guarantees of Google’s Privacy Sandbox APIs—Private Aggregation API (PAA) and Attribution Reporting API (ARA)—under differential privacy (DP) in the realistic, challenging setting of *interactive queries* and *dynamic database updates*. We first construct an abstract semantic model of the APIs, explicitly formalizing query strategies and database evolution as response-dependent processes. Under reasonable assumptions, we rigorously prove that the joint mechanism formed by PAA and ARA satisfies $(varepsilon,delta)$-differential privacy. Our analysis breaks from classical DP frameworks, which typically assume static databases and non-interactive queries. This is the first formal verification establishing end-to-end DP guarantees for Privacy Sandbox’s advertising measurement infrastructure. By bridging theory and practice, our work provides a verifiable, mathematically grounded privacy assurance for industrial-scale privacy-preserving APIs—filling a critical gap in the formal verification of real-world privacy mechanisms.

Technology Category

• Machine Learning: Privacy
• Application Domains: Mobility, Driving & Flight
• Multiagent Systems: Mechanism Design

Application Category

• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems

Abstract

The Privacy Sandbox initiative from Google includes APIs for enabling privacy-preserving advertising functionalities as part of the effort around limiting third-party cookies. In particular, the Private Aggregation API (PAA) and the Attribution Reporting API (ARA) can be used for ad measurement while providing different guardrails for safeguarding user privacy, including a framework for satisfying differential privacy (DP). In this work, we provide an abstract model for analyzing the privacy of these APIs and show that they satisfy a formal DP guarantee under certain assumptions. Our analysis handles the case where both the queries and database can change interactively based on previous responses from the API.

Problem

• Analyzing privacy of Privacy Sandbox APIs
• Ensuring differential privacy in ad measurement
• Handling interactive queries and database changes

Innovation

• APIs enable privacy-preserving advertising functionalities
• Private Aggregation API ensures differential privacy
• Interactive query and database changes analyzed