Evolving medical AI research from point-in-time diagnostics to long-term disease management requires a strong commitment to scientific rigor. Published today in Nature, our latest research with Google DeepMind introduces a new milestone for AMIE. In a randomized, blinded study using multi-visit scenarios, AMIE’s multi-agent architecture demonstrated longitudinal management reasoning on par with 21 primary care physicians. Read the full breakdown of how we're exploring the future of assistive medical AI below 👇
New publication from Google Research and Google DeepMind in Nature: We advance AMIE, our research medical AI, from one-off diagnostic conversations toward treating and managing disease over time, using clinical guidelines and drug formularies. In our randomized study, now published in Nature, we showed that AMIE demonstrated physician-level capabilities for treating and managing disease over sequential, multi-visit encounters with patient actors. The findings suggest that conversational medical AI systems like AMIE could one day help augment care and give doctors back time with their patients where it truly matters. While AMIE remains an experimental research system and is not ready for medical deployment, these peer-reviewed findings contribute to our strong commitment to evidence generation in AI for medical applications, offer an early look at how AI could help support long-term disease management. Learn more: https://goo.gle/3SaNXDW