Trusted research environments (TREs) provide secure access to very sensitive data for research. All TREs operate manual checks on outputs to ensure there is no disclosure risk. Machine learning (ML) models require a large amount of data. If the data is personal, the TRE is a well-established data management solution. However, ML models present disclosure risks in both type and scale. 

ML use in TREs create a risk of releasing raw or disclosive data from the safe environment. Once ML models are in the world, uses of the algorithm are uncontrolled, there is not enough overall models on their capability, who can use them once outside TREs, and how they could be used to reidentify individuals within the data.