October 27, 2026

Microsoft says its AI shows ‘higher diagnostic accuracy’ than doctors in tests, study finds

The model scores above 80 percent in diagnostic tests, but its real-world performance remains uncertain

Microsoft says its AI shows ‘higher diagnostic accuracy’ than doctors in tests, study finds
An expanded view of Surface Laptop Ultra stands on display at Microsoft Windows & Surface event where new on-device AI capabilities for Windows PCs using Nvidia's new RTX Spark chips were announced, in San Francisco, California, US, Oct 7, 2026. REUTERS/Manuel Orbegozo

bdnews24.com News Service

Published : 26 Oct 2026, 11:54 PM

Updated : 26 Oct 2026, 11:54 PM

Microsoft says its new artificial intelligence system can diagnose patients more accurately than generalist doctors in controlled tests, but questions remain over how it would perform in real healthcare settings, Futurism reports.

The system, called Microsoft AI Diagnostic Orchestrator (MAI-DxO), was assessed in a peer-reviewed study published in the journal Science.

Researchers tested it on 304 medical case reports from the New England Journal of Medicine and 325 emergency department cases from the MIMIC-IV-ED database.

Futurism notes that MAI-DxO scored 83.2 percent accuracy on medical cases and 80.3 percent on emergency cases, compared with doctors’ 36.1 percent and 46.8 percent respectively.

The model also reached diagnoses with estimated testing costs half to one-third those of human doctors, cited per the report.

Microsoft AI chief Mustafa Suleyman told The Verge that major deployments were expected in 2027, saying the technology would “change the way care is delivered in a big way”, according to Futurism.

However, Futurism explains that the research took place in a controlled benchmark setting, so the results do not establish how well the system would work in everyday medical practice.

The AI worked through cases sequentially, asking follow-up questions and selecting tests, but relied on another AI agent to retrieve details from existing records, limiting how closely it replicated doctor-patient interactions.

The study highlights the AI’s diagnostic potential but notes that real-world care also depends on patient communication and details absent from prepared case files.

Related Stories