A technique that adds statistical noise to data to preserve individual privacy while allowing useful aggregate insights. Can also be used when augmenting real data with synthetic data to create larger data sets for testing and validation.