[Paper Review] Privacy Risk and Preservation For COVID-19 Contact Tracing Apps

Dong Wang, Fang Liu|arXiv (Cornell University)|Jun 27, 2020

COVID-19 Digital Contact Tracing4 citations

TL;DR

This paper analyzes privacy risks and preservation mechanisms in GPS- and Bluetooth-based contact-tracing apps deployed during the COVID-19 pandemic, comparing centralized and decentralized architectures. It evaluates the effectiveness of privacy-preserving techniques such as local data processing, anonymization, and cryptographic protocols, concluding that decentralized models with on-device computation significantly reduce exposure to surveillance and data misuse.

ABSTRACT

Contact tracing in the COVID-19 pandemic is key to prevent the further spread of COVID-19. Countries and regions around the world have developed and deployed or are considering adopting contact-tracing software or mobile apps. While contact tracing apps and software play an important role in the pandemic, red flags have been raised regarding the privacy risk associated with contact tracing. In this short paper, we provide an overview on the GPS and Bluetooth based contact-tracing apps in the framework of both centralized and decentralized models, examine the associated privacy risk and the effectiveness of the privacy-preserving measures adopted in different apps.

Motivation & Objective

• To assess the privacy risks associated with national contact-tracing apps developed during the COVID-19 pandemic.
• To evaluate the effectiveness of privacy-preserving mechanisms in existing contact-tracing applications.
• To compare the security and privacy trade-offs between centralized and decentralized contact-tracing architectures.
• To identify systemic vulnerabilities in data collection, storage, and transmission across different app designs.
• To provide guidance on improving privacy-by-design principles in public health technology deployments.

Proposed method

• Categorizing contact-tracing apps based on their underlying technology: GPS-based versus Bluetooth-based.
• Classifying systems into centralized and decentralized models based on data storage and processing location.
• Analyzing the role of cryptographic techniques such as ephemeral identifiers and local key management in preserving user anonymity.
• Evaluating the impact of data minimization and on-device processing in reducing the attack surface for privacy breaches.
• Assessing the transparency and auditability of systems through public documentation and open-source availability.
• Comparing real-world implementations using criteria such as data retention policies, access controls, and third-party data sharing.

Experimental results

Research questions

• RQ1 What are the primary privacy risks associated with centralized versus decentralized contact-tracing architectures?
• RQ2 How effective are on-device processing and anonymization techniques in preventing user identification?
• RQ3 To what extent do existing contact-tracing apps implement end-to-end privacy-preserving mechanisms?
• RQ4 What are the implications of data retention and access policies on long-term user privacy?
• RQ5 How do technical design choices influence the balance between public health efficacy and individual privacy?

Key findings

• Decentralized models that store exposure logs locally on users' devices significantly reduce the risk of mass surveillance and data breaches.
• Apps using Bluetooth-based proximity detection with rotating identifiers demonstrated stronger privacy guarantees than GPS-based alternatives.
• Centralized systems were found to be more vulnerable to unauthorized access and government overreach due to centralized data repositories.
• The use of end-to-end encryption and local data processing substantially diminished the likelihood of user re-identification.
• Lack of transparency and absence of open-source code in some apps raised concerns about hidden data collection and misuse.
• Even with privacy-preserving features, some apps still shared metadata with third parties or retained data beyond necessary durations, undermining trust.