This article analyzes the technical nuances and legal implications of using synthetic data, which uses machine learning techniques to modify raw data, as an alternative to anonymization or differential privacy to protect privacy interests while facilitating research.