On the legal nature of synthetic data

The present manuscript attempts to analyze the legal qualification of synthetic data generated from personal data. Three main conclusions are drawn from our legal analysis: first, full data protection compliance prior to data synthesis would be applicable in many cases; second, according to the identifiability test as enshrined in the definition of personal data, synthetic data will be considered pseudonymous or anonymous data depending on the appropriateness of the data synthesis and the related ex-post control mechanisms; third, the question of legal qualification remains, however, an unresolved issue in light of the exegetical discrepancy of the identifiability test and the dissonance over the data protection model required by law and doctrine.