In Chap. 1, I presented and discussed the state-of-the-art ML applications in medical diagnosis and the opportunities they have to bring about beneficial outcomes to patients, clinicians and perhaps even to the healthcare systems at large. This was done from a medical and socio-technical perspective grounded in the actual (in the sense of both current and concrete) technical possibilities of the models and the headways made in terms of their development and deployment. Research evidence showed that some ML models in diagnosis meet and sometimes exceed benchmarks that rival those of human doctors and seemed to suggest that the models could, at the least, assist them in the diagnostic process.