Towards Expert-level Medical AI for Real-time Video Consultations
AMIE’s video version matched or beat primary-care doctors in a simulated telehealth exam, while still trailing them on rapport.
The paper tests a Gemini-based multi-agent medical AI in randomized OSCE video consultations with 30 primary care physicians, 15 patient actors, and 100 scenarios. Clinical evaluators rated AMIE (Video) on par with or better than doctors for history-taking, diagnosis, management, and physical observation. Patient actors preferred AMIE for assessing and explaining conditions, but preferred physicians for partnership building. The authors say limits remain in anatomical precision, subtle affect, and fast movements, and that more research is needed before real-world use. ArXiv · AI/CL/LG's note
The paper tests a Gemini-based multi-agent medical AI in randomized OSCE video consultations with 30 primary care physicians, 15 patient actors, and 100 scenarios. Clinical evaluators rated AMIE (Video) on par with or better than doctors for history-taking, diagnosis, management, and physical observation. Patient actors preferred AMIE for assessing and explaining conditions, but preferred physicians for partnership building. The authors say limits remain in anatomical precision, subtle affect, and fast movements, and that more research is needed before real-world use. ArXiv · AI/CL/LG's note
score 7