Microsoft says its AI tool is better than doctors at diagnosing conditions

Microsoft announced Thursday that it has built an AI tool it claims can diagnose medical conditions up to four times more accurately than human doctors, according to results published in the journal Science. Mustafa Suleyman, the company's AI chief, told The Verge that healthcare AI "is finally here."
The evaluation pitted Microsoft's system against human physicians across scenarios assembled from medical case reports and emergency department data. Microsoft gathered those cases into a benchmark of its own construction for the comparison, rather than relying on an existing public dataset.
That detail is worth noting for anyone weighing the claim. A purpose-built benchmark lets the developers curate which cases are included, and the headline multiplier reflects performance on that specific set of scenarios rather than clinical practice. The published paper in Science gives outside researchers the chance to inspect the methodology, but the four-times figure should be read as a result on Microsoft's benchmark, not a measure of diagnostic accuracy in a real hospital.
Why it matters: If the results hold up under external scrutiny, diagnostic AI moves closer to deployment in triage and clinical decision support — a shift developers building health products will need to account for, and one regulators and clinicians will scrutinize closely. For now, the claim is Microsoft's own, and independent replication is the test that counts.
Based on reporting from the original publisher. Visit the source for full context and later updates.
Publisher excerpt
For years, AI leaders have touted healthcare as one of the clearest examples of where their technology will improve lives. Now "it is finally here," Microsoft AI CEO Mustafa Suleyman tells The Verge. The company claims to have developed an AI-powered tool it says can diagnose conditions up to four times more accurately than human […]