The uses for this kind of tool in diagnostics is far-reaching but there are a couple worth noting, such as the FDNA system which uses facial recognition software to screen patients for over 8000 diseases and rare genetic disorders with a scarily high degree of accuracy. Or the New York University team that created an Al capable of scanning thousands of medical documents to pinpoint patients at risk of developing diabetes, heart failure, or stroke. So far it has never been wrong. Clearly robots are just better at giving a diagnosis.