In an interview with the Financial Times, the chief executive of Microsoft AI said the trial was a step on the path to \u201cmedical superintelligence\u201d that could help solve staffing crises and long waiting times for overstretched health systems. <\/p>\n
Microsoft\u2019s new system is underpinned by a so-called \u201corchestrator\u201d that creates virtual panels of five AI agents acting as \u201cdoctors\u201d \u2014 each with a distinct role, such as coming up with hypotheses or choosing diagnostic tests \u2014 which interact and \u201cdebate\u201d together to choose a course of action.<\/p>\n
To test its capabilities, \u201cMAI-DxO\u201d was fed 304 studies from the New England Journal of Medicine (NEJM) that describe how some of the most complicated cases were solved by doctors.\u00a0<\/p>\n
This allowed researchers to test if the programme could figure out the correct diagnosis and relay its decision-making process, using a new technique called \u201cchain of debate\u201d, which makes AI reasoning models give a step-by-step account of how they solve problems.<\/p>\n
Recommended<\/p>\n
Microsoft used leading large language models from OpenAI, Meta, Anthropic, Google, xAI and DeepSeek. The orchestrator made all LLMs perform better, but worked best with OpenAI\u2019s o3 reasoning model to correctly solve 85.5 per cent of the NEJM cases.<\/p>\n That compared with about 20 per cent by experienced human doctors, but those physicians were not allowed access to textbooks or to ask colleagues in the trial, which could have increased their success rate.<\/p>\n A version of the technology could soon also be deployed in Microsoft\u2019s Copilot AI chatbot and Bing search engine, which handle 50mn health queries a day.<\/p>\n Suleyman said Microsoft is nearing \u201cAI models that are not just a little bit better, but dramatically better, than human performance: faster, cheaper and four times more accurate\u201d.<\/p>\n \u201cThat is going to be truly transformative,\u201d he added.<\/p>\n Suleyman\u2019s new effort comes after Deepmind has led the way on AI-related<\/a> heathcare breakthroughs. The Google lab\u2019s chief Sir Demis Hassabis jointly won a chemistry Nobel Prize last year for using AI to unlock the biological secrets of proteins that underpin life.\u00a0<\/p>\n Microsoft has invested almost $14bn into OpenAI and has exclusive rights to use and sell its technology. However, the tech giant is embroiled in high-stakes brinkmanship<\/a> with the start-up, which is attempting to convert into a for-profit entity, with both sides clashing over the future terms of their partnership.<\/p>\n Suleyman said that while OpenAI\u2019s model performed the best, Microsoft was \u201cagnostic\u201d over which of the four \u201cworld-class models\u201d MAI-DxO used.\u00a0<\/p>\n \u201cWe have long believed that they\u2019ll become commodities\u2009.\u2009.\u2009.\u2009it\u2019s the aggregate orchestrator which I think is the differentiator,\u201d he said.<\/p>\n Recommended<\/p>\n![\"\"\/](\"https:\/\/www.europesays.com\/uk\/wp-content\/uploads\/2025\/06\/https:\/\/www.ft.com\/__origami\/service\/image\/v2\/images\/raw\/https%3A%2F%2Fd1e00ek4ebabms.cloudfront.net.jpeg\")
<\/a><\/p>\n