Google's AMIE conversational AI shows strong performance in real-world primary care study
A prospective feasibility study published in The Lancet found AMIE AI conducted pre-visit chats with 98 real primary care patients, showing 90% inclusion of final diagnoses (75% top-3 accuracy), zero safety interventions, and high quality ratings.
TLDR
This represents one of the first real-world, prospective tests of conversational diagnostic AI in clinical practice, moving beyond simulations and benchmarks. Results suggest potential for AI to assist physician preparation and improve access to care, though findings also highlight current limitations and underscore the need for supervised integration and further validation before clinical deployment.
Combined views
29.2K
1 Source, first seen ago
