Microsoft's New Medical AI Model Outperforms Doctors in Diagnosis

Microsoft announced that its new medical artificial intelligence model, named Microsoft AI Diagnostic Orchestrator (MAI-DxO), demonstrated higher diagnostic accuracy than general practitioners in controlled studies.

The medical artificial intelligence model developed by Microsoft, named MAI-DxO, was the subject of a peer-reviewed study published in the journal Science. The research revealed that the artificial intelligence approaches each case like a real doctor, following a process different from traditional testing methods.

Within the scope of the study, the model was examined on 304 clinicopathological conference cases from the New England Journal of Medicine and 325 emergency department admission cases from the MIMIC-IV-ED database. In the first group, the artificial intelligence achieved an accuracy of 83.2 percent, while general practitioners remained at 36.1 percent. In emergency department cases, the artificial intelligence reached an accuracy rate of 80.3 percent, and human doctors reached 46.8 percent. Additionally, Microsoft reported that the artificial intelligence performed test costs at half or a third of the cost of human doctors. It is noted that its success in a real medical environment is debatable due to the tests being conducted in a controlled benchmarking environment.