Microsoft AI diagnostic system beats doctors on hard cases in Science study

Microsoft published a study in Science saying its AI diagnostic system diagnosed conditions up to four times more accurately than human physicians. Tests used scenarios from medical case reports and emergency department data, compiled into Microsoft's own benchmark. Reporting notes most of the gain appears to come from the underlying AI models rather than Microsoft's system.

Why it matters: Peer-reviewed publication in Science gives this more weight than a vendor blog claim, but the self-built benchmark and the finding that the base models drive most gains limit what it proves. It suggests medical performance is tracking general model progress, so improvements will arrive broadly, while real-world clinical validation remains the open question.