Privacy-Enhancing Technologies (PETs)

Definition

Technologies that support data processing while minimizing personal information exposure through reduced identifiability, information concealment and data splitting, balancing privacy protection with data use.

Data use

Privacy protection

PETPrivacy-enhancing technologies (umbrella term)

Reduced identifiability · Derived data generation

Synthetic data
Differential privacy

Concealment · Shielding

Homomorphic encryption
Zero-knowledge proofs

Data splitting · Access control

Multiparty computation (MPC)
Federated learning

Technologies can be combined · Protection depends on how they are applied

Source: Financial Security Institute (2025-11-24), PET classification

PET characteristics and implementation considerations

Privacy alongside data use: PETs support statistics, joint computation and AI training while reducing exposure of personal information.
Minimising exposure: PETs reduce exposure by lowering identifiability or concealing information. Other approaches split data for computation or keep original data with each organisation during processing.
An umbrella term: PETs encompass multiple technologies with different protection mechanisms.
Selection by purpose: Consider protection, analytical accuracy and processing costs for the intended use. With differential privacy, for example, the noise added can affect analytical accuracy.
Technologies can be combined: One technology can complement another. Federated learning, for example, can be combined with differential privacy.
Protection depends on implementation: PETs do not automatically guarantee complete anonymity or prevent all leaks. Protection varies with the technology and its settings; re-identification risk needs ongoing assessment.