Ethical Data Economy

The thesis presents a SHA-256 Merkle-tree contribution system aligned with the FACT principles (Fairness, Accuracy, Confidentiality, Transparency) introduced by Sohail et al. as a successor to FAIR. The architecture is inspired by the decentralized Wibson protocol but redesigned around a centralized Governor model that fits a single-company booking service.

Three roles: the User (data producer and owner, rewarded per contribution), the Service Provider (the buyer of aggregated signals — never the individual), and the Governor (Fixmeapp — verifies signal validity, signs matches, and intermediates the transaction). Raw user data is pseudo-anonymized within 30 days (demographic data as buckets, locations as geographical polygons) and then encoded in a Merkle tree using SHA-256 hashing. Only the trees — cryptographically resistant to preimage, second-preimage and collision attacks — are queryable from outside.

The proof-of-concept successfully maps 10,000 synthetic users into provable, queryable trees in roughly one second per thousand entries. It is evaluated against quantitative FACT metrics: discrete Gini coefficient for fairness, Dawid-Skene "no-truth" accuracy, ε-differential privacy for confidentiality, and an effectiveness measure for transparency. Structural confidentiality is guaranteed by the hash architecture; the conclusion validates feasibility at scale.

This thesis builds a predicate-based verification system inspired by the WibsonTree Protocol. The prototype uses Ordered Binary Decision Diagrams (OBDD) that decouple raw data storage from eligibility verification — allowing the platform to verify user attributes "blindly" without ever loading the underlying data.

The system is benchmarked in a controlled environment against two alternatives: a traditional raw-storage baseline and a Zero-Knowledge Proof system using the Groth16 proving scheme. Findings: predicate-based architectures offer an efficiency-privacy compromise — lighter than ZKP (which has prohibitive computational costs), much safer than raw storage. The privacy tax: approximately 52.5× larger communication payload than the traditional approach. Verifications process at acceptable frequency without critical bottlenecks.

The conclusion is direct: "the storage of raw sensitive data is a legacy practice rather than a necessity in centralized social platforms." Verification of user attributes can be decoupled from storage, and the architecture is practical at scale.