Preserving privacy in European health research: The case of synthetic data

Abstract

This paper investigates the role of synthetic data in the field of health research, with a particular focus on data protection. More specifically, it aims at clarifying whether this new technology represents an alternative to more classic anonymisation techniques. The analysis is construed on a review of the existing literature; nevertheless, it is noted that the majority of contributions focuses on the technical aspects of synthetic data and machine learning, while less legal studies have been conducted on this topic. The outcome of this study outlines that, by using synthetic data which respects the ‘privacy by design’ principle (although the identifiability risk still exists), researchers are no longer occupied by the question of re-identification but rather focus on the quality and utility of synthetic datasets. After examining the different solutions applied to enshrine privacy, however, this paper concludes there is a necessity for regulating the use of artificially generated data for research and machine learning purposes.