Privacy enhancing technologies

What it covers

Data-driven systems and artificial intelligence (AI) increasingly rely on large volumes of personal and sensitive data. As these systems enable profiling, automated decision-making and new forms of inference, privacy risks can arise even when identifying information has been removed. This unit examines what privacy means in technical and societal contexts, and develops practical skills in selecting, applying and evaluating privacy-enhancing technologies. Topics may include privacy requirements and threat models; differential privacy (DP) for data analysis; secure multiparty computation (MPC), zero-knowledge proofs (ZKP) and fully homomorphic encryption (FHE) for computation over protected data; and post-quantum cryptography (PQC) for security against quantum-enabled adversaries. These techniques have a wide range of applications including in AI-based systems and digital ID technologies. The unit examines the guarantees, assumptions, implementation challenges and privacy-utility-performance trade-offs associated with these technologies. The unit will help the participants design and evaluate solutions that enable beneficial uses of data while limiting disclosure and supporting the responsible development of digital and AI systems.

Assessment

When and where