Quantum-Secure Data Pseudonymization Framework with QKD-based Key Management and ML-Driven Re-Identification Risk Analysis
The increasing needs in data sharing in the fields of finance, governance, and artificial intelligence pose a major threat to privacy, particularly in quantum computing. In this paper, a hybrid privacy-preserving system incorporating simulated BB84 Quantum Key Distribution (QKD) to generate secure keys, reversible pseudonymization with encrypted mapping vaults, automatic key rotation, and re-identification risk analysis by machine learning are introduced. Also, optional differential privacy layer allows irreversible anonymization in the cases of analysis. The proposed system will enable two modes, that is, recovery of secure data and the ability to publish data in privacy modes. The experimental findings indicate that authorized users have 100% recovery accuracy, re-identification risk is low and is close to random guessing and data utility is acceptable given the privacy restrictions. FastAPI and Streamlit are used to implement the framework, which is appropriate in the real-world deployment in clouds.