Privacy-Enhancing Technologies (PETs)

Technical and organisational measures (e.g., differential privacy, secure multi-party computation, federated learning, homomorphic encryption) designed to minimise privacy risk while enabling useful data analysis and collaboration. In the NICPET context PETs are the technical approaches evaluated, piloted and operationalised to enable privacy-preserving public-sector data use.

Technical measures and techniques that minimize the use or disclosure of personal data, maximize data security, and help preserve data subject privacy during AI development and operation; examples include differential privacy, federated learning, homomorphic encryption, synthetic data, anonymization, and pseudonymization.