Microsoft's artificial intelligence surpasses doctors in diagnosing complex diseases

Microsoft has developed an artificial intelligence (AI) system for medical diagnostics that can assist or even replace doctors in solving complex cases. The tool, named Microsoft AI Diagnostic Orchestrator (MAI-DxO), was created by the Microsoft AI division under the leadership of Mustafa Suleyman, co-founder of DeepMind. MAI-DxO includes a virtual consult comprising five AI agents who act as doctors of various specialties. These agents interact, discuss options, and together provide a conclusion. The system was tested on 304 complex clinical cases from the New England Journal of Medicine, and the results were very impressive. The Financial Times reported that 'the agents interact, discuss options, and collectively formulate a conclusion.' The testing used a new methodology called 'debate chain,' which ensures transparency in the algorithm's logic. The most effective model in testing was o3 from OpenAI, which correctly solved 85.5% of medical tasks, significantly saving costs. The system can be integrated into Microsoft products that handle a massive amount of medical requests daily. Company employees noted that their goal is to create an AI that surpasses humans in speed, accuracy, and cost-effectiveness. Mustafa Suleyman stated, 'We are approaching the creation of AI that will significantly surpass humans in speed, accuracy, and cost-effectiveness.' Dominic King, former head of DeepMind and now an employee at Microsoft, highlighted the potential efficiency of the new system in healthcare but called for additional testing before widespread use. A successful heart transplant operation using a robot was also recorded in the U.S. at Baylor St. Luke's Medical Center, where a 45-year-old man with heart failure received a new organ.