Abstract
Deep learning techniques and whole-genome sequencing promise to increase well-being but also risk perpetuating psychological essentialism, potentially justifying inequality. In this Comment, we offer two much-needed systematic frameworks for clinicians and researchers to avoid essentialist inferences and unfair treatment: (1) a data-driven method for detecting causal fairness in precision health and (2) an ethical framework for determining when it is morally permissible to use racial classifications in population health research.