Tech giant Microsoft, recently hit with a fresh round of layoffs, has developed a new medical AI tool that performs better than human doctors at complex health diagnoses, creating a “path to medical superintelligence”. The Microsoft AI team shared research that demonstrated how AI can sequentially investigate and solve medicine’s most complex diagnostic challenges—cases that expert physicians struggle to answer.
Tech company’s AI unit, led by the British tech pioneer Mustafa Suleyman, has developed a system that imitates a panel of expert physicians tackling “diagnostically complex and intellectually demanding” cases.
Microsoft AI Diagnostic Orchestrator (MAI-DxO) correctly diagnosed up to 85% of NEJM case proceedings, a rate more than four times higher than a group of experienced physicians. MAI-DxO also gets to the correct diagnosis more cost-effectively than physicians, the company said in a blog post.
Microsoft says AI system better than doctors
The Microsoft AI Diagnostic Orchestrator”, or MAI-DxO for short, the AI-powered tool is developed by the company’s AI health unit, which was founded last year by Mustafa Suleyman.
The tech giant said when paired with OpenAI’s advanced o3 AI model, its approach “solved” more than eight of 10 case studies specially chosen for the diagnostic challenge. When those case studies were tried on practising physicians – who had no access to colleagues, textbooks or chatbots – the accuracy rate was two out of 10. Microsoft said it was also a cheaper option than using human doctors because it was more efficient at ordering tests.
When benchmarked against real-world case records, the new medical AI tool “correctly diagnoses up to 85% of NEJM case proceedings, a rate more than four times higher than a group of experienced physicians” while being more cost-effective.
What’s impressive is that these cases are from the New England Journal of Medicine and are very complex and require multiple specialists and tests before doctors can reach any conclusion.
According to The Wired, the Microsoft team used 304 case studies sourced from the New England Journal of Medicine to devise a test called the Sequential Diagnosis Benchmark. A language model broke down each case into a step-by-step process that a doctor would perform in order to reach a diagnosis.
Microsoft new AI tool diagnosed 85% cases
For this, the company used different large language models from OpenAI, Meta, Anthropic, Google, xAI and DeepSeek. Microsoft said that the new AI medical tool correctly diagnosed 85.5 per cent of cases, which is way better compared to experienced human doctors, who were able to correctly diagnose only 20 per cent of the cases.
“This orchestration mechanism—multiple agents that work together in this chain-of-debate style—that’s what’s going to drive us closer to medical superintelligence,” Suleyman told The Wired.
Microsoft announced it is building a system designed to mimic the step-by-step approach of real-world clinicians—asking targeted questions, ordering diagnostic tests, and narrowing down possibilities to reach an accurate diagnosis. For example, a patient presenting with a cough and fever might be guided through blood tests and a chest X-ray before the system determines a diagnosis like pneumonia.
Microsoft said its approach was able to wield a “breadth and depth of expertise” that went beyond individual physicians because it could span multiple medical disciplines.
It added: “Scaling this level of reasoning – and beyond – has the potential to reshape healthcare. AI could empower patients to self-manage routine aspects of care and equip clinicians with advanced decision support for complex cases.”
Microsoft acknowledged its work is not ready for clinical use. Further testing is needed on its “orchestrator” to assess its performance on more common symptoms, for instance.
