TL;DR
This paper introduces Kovacs, a novel system that leverages permissionless blockchains to achieve decentralized inverse transparency for employee data. It tackles the challenge of tracking data usage while ensuring privacy and GDPR compliance, offering a secure and tamper-proof log without relying on trusted third parties.
What Problem Does This Paper Address?
In today's data-driven workplaces, "people analytics" tools use employee data for various purposes, from improving collaboration to making predictions. However, this raises significant risks of data misusage and profiling, especially given the inherent power asymmetry between employers and employees. Inverse transparency aims to mitigate these risks by providing employees with a way to track and monitor how their personal data is used.
Existing solutions for inverse transparency are often centralized, requiring trust in multiple parties like employers and system administrators. This creates a single point of failure and potential for manipulation. Furthermore, the immutability of blockchain, while beneficial for tamper-proofing, directly conflicts with privacy regulations like GDPR, which mandate the right to rectification and erasure of personal data. This paper addresses the critical need for a decentralized, secure, and GDPR-compliant system for inverse transparency.
Methods
Kovacs is built upon two core components:
- Non-repudiable Data Exchange Protocol: This protocol ensures that data exchanges are provable and cannot be denied by either party. It's based on established non-repudiation protocols but adapted for a decentralized, blockchain-based environment.
- Private Pseudonym Provisioning: This algorithm generates unique, one-time pseudonyms for data owners and consumers. These pseudonyms are designed to be unlinkable to real identities, crucial for GDPR compliance (anonymity), while still allowing proof of ownership for exercising data rights.
The system uses a permissionless blockchain to store usage logs immutably. Each log entry is encrypted using one-time pseudonyms, ensuring confidentiality. The system also incorporates a "fake chatter" protocol to mask real data exchanges with fabricated traffic, further enhancing privacy.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Blockchain as a Solution for Decentralized Trust: The paper highlights how permissionless blockchains, with their inherent immutability and decentralization, are well-suited for creating tamper-proof logs without relying on a single trusted entity.
- Bridging the GDPR-Blockchain Divide: A key insight is that the conflict between blockchain's immutability and GDPR's right to erasure can be resolved through pseudonymization and careful management of the link between pseudonyms and real identities off-chain.
- Pseudonymity as a Privacy Enabler: The use of one-time, unlinkable pseudonyms is central to achieving both anonymity (for GDPR) and accountability (for inverse transparency).
- The "Fake Chatter" Protocol: This novel approach to masking real network activity with fabricated traffic is a clever way to enhance privacy in a system where communication patterns could otherwise reveal sensitive information.
Key Results
- Security: Kovacs is shown to be robust against various attacks, including repudiation, fabrication, and identity derivation, due to its cryptographic foundations and protocol design.
- GDPR Compliance: The system effectively addresses the right to erasure by ensuring that the link between pseudonyms and real identities is managed off-chain and can be deleted, thereby anonymizing the data stored on the blockchain.
- Performance: While not optimized for speed, Kovacs demonstrates manageable exchange durations (median of 6.4s for data owner, 10.6s for data consumer) and blockchain append times (median of 2.3s). The fake chatter protocol increases data consumer exchange duration to a median of 25.0s.
- Scalability: Log retrieval time scales linearly with the number of stored logs, and storage requirements are also manageable, increasing linearly with log entries.
Limitations and Future Work
- Performance Trade-off: The paper acknowledges that the focus on security and decentralization comes at the cost of exchange speed. For scenarios where performance is paramount and security risks are lower, less secure solutions might be preferable.
- Institutional IdP Dependency: The implemented identity verification relies on an institutional Identity Provider (IdP), which is a trusted third party. The authors suggest that a fully decentralized solution would require a web-of-trust model.
- Legal Analysis: The paper provides a technical analysis of GDPR compliance but notes that a formal legal analysis for specific use cases is still necessary.
- User-Side Security: The system's security can still be compromised if individual users' devices are hacked or if they are tricked into revealing sensitive information.
- Future Work: The authors suggest exploring alternative identity verification schemes (like web-of-trust) for fully decentralized inverse transparency and conducting real-world evaluations to assess performance in diverse scenarios. They also highlight the need for broader analysis considering user and operator perspectives.
Summary updated