[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-microsoft-says-diagnostic-ai-beats-doctors-by-up-to-4x":10,"sections":40},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":30,"tags":31,"sources":35,"feedback":39,"feedback_at":22,"cost_usd":39,"total_tokens":39},12184,"microsoft-says-diagnostic-ai-beats-doctors-by-up-to-4x","Microsoft Says Diagnostic AI Beats Doctors by Up to 4x","Microsoft says its AI beat physicians by up to four times on diagnoses, in a Science paper built on a benchmark the company compiled itself.","Microsoft says it has built an AI diagnostic tool that is up to four times more accurate than human doctors, and the results are now published in the journal Science.\n\nAccording to the company, the tests pitted its system against human physicians on a series of scenarios drawn from medical case reports and emergency department data. Microsoft compiled those cases into its own benchmark, a fixed set of problems used to score the tool. Microsoft AI CEO Mustafa Suleyman said of AI's long-promised payoff in healthcare that \"it is finally here.\"\n\nThat is a bigger claim than the usual chatbot demo. AI leaders have pointed to healthcare for years as the clearest case for their technology, and a peer-reviewed paper gives this result more weight than a blog post would.\n\nStill, two details deserve attention. First, \"up to four times\" describes a best result, not necessarily the typical gap between the tool and the doctors. Second, peer review checks a study's method, but it does not turn a company-built benchmark into independent validation. The test and the tool come from the same place.\n\nCase reports also differ from a live clinic. They are written up after the fact, with the relevant findings already gathered, while real patients arrive with incomplete histories and messy data. A tool that excels on tidy cases may not keep its edge at a bedside.\n\nThe closest comparison is a drugmaker's own trial versus one run by an outside lab. Both are useful, but only the second settles the question. The open issue is whether a different group, using cases Microsoft did not select, would find the same gap.","[\"microsoft\",\"healthcare-ai\",\"diagnostics\"]","2026-10-22T18:00:00.000Z","2026-10-22T18:36:37.077Z","2026-10-22T18:36:37.347Z","published",null,[24],{"id":25,"reviewer":26,"round":27,"reason":28,"status":29},"editor-r1","editor",1,"The lede's 'published October 22' date is not in the source, which says only 'Thursday', so use the sourced wording or drop it; also replace the reader-facing 'the source does not say' phrasing with attribution such as 'The Verge's report doesn't say' or 'Microsoft hasn't disclosed', and soften the dek's 'backs' since publication in Science doesn't independently validate a company-built benchmark.","resolved","ai",[32,33,34],"microsoft","healthcare-ai","diagnostics",[36],{"name":37,"url":38},"The Verge","https:\u002F\u002Fwww.theverge.com\u002Fai-artificial-intelligence\u002F1016690\u002Fsuleyman-microsoft-diagnostic-orchestrator-science-medical-superintelligence",0,{"sections":41},[42,45,50,55,60,65,70,75,80,85,90,95,100,105],{"name":43,"slug":30,"count":44,"latest_published_at":18},"AI",7064,{"name":46,"slug":47,"count":48,"latest_published_at":49},"Security","security",1051,"2026-10-22T15:11:26.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Policy","policy",571,"2026-10-22T16:00:29.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Deals","deals",558,"2026-10-22T13:00:00.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":64},"Hardware","hardware",288,"2026-10-22T17:25:00.000Z",{"name":66,"slug":67,"count":68,"latest_published_at":69},"Consumer Tech","consumer-tech",282,"2026-10-22T17:42:52.000Z",{"name":71,"slug":72,"count":73,"latest_published_at":74},"Science","science",215,"2026-10-22T16:37:00.000Z",{"name":76,"slug":77,"count":78,"latest_published_at":79},"Startups","startups",142,"2026-10-22T16:05:08.000Z",{"name":81,"slug":82,"count":83,"latest_published_at":84},"Software","software",140,"2026-10-22T18:01:29.000Z",{"name":86,"slug":87,"count":88,"latest_published_at":89},"Dev Tools","dev-tools",122,"2026-10-21T16:00:00.000Z",{"name":91,"slug":92,"count":93,"latest_published_at":94},"General","general",96,"2026-10-22T16:00:00.000Z",{"name":96,"slug":97,"count":98,"latest_published_at":99},"Gaming","gaming",86,"2026-10-22T13:17:59.000Z",{"name":101,"slug":102,"count":103,"latest_published_at":104},"Reviews","reviews",38,"2026-10-20T12:00:00.000Z",{"name":106,"slug":107,"count":108,"latest_published_at":109},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]