AMIE, our research medical AI system, demonstrates real-time clinical video consultation capabilities in a first-of-its-kind study.

Google Research and Google DeepMind have developed AMIE, a research medical AI system, to conduct real-time clinical video consultations. Built on Gemini and Project Astra with a multi-agent architecture, AMIE interprets visual and auditory cues, guides virtual physical exams, and provides real-time diagnostic reasoning. The system was evaluated in a randomized study using simulated consultations with patient actors and primary care physicians, assessing core clinical competencies such as history-taking, diagnostic accuracy, and communication quality.

In the study, clinical evaluators found AMIE performed favorably across multiple clinical competencies, including thoroughness in history-taking, diagnostic accuracy, appropriateness of management, and quality of communication. Patient actors involved in the simulated consultations also expressed a preference for the video-based interaction over text-based chat. The research highlights AMIE's ability to process and respond to non-verbal cues, a key aspect of traditional clinical consultations.

AMIE remains a research system, and the team emphasizes that further research is required before any real-world clinical deployment. The study represents a first-of-its-kind demonstration of an AI system capable of conducting real-time clinical video consultations at an expert level. The findings suggest potential for future applications in telemedicine and remote patient care.

The research was conducted by Google Research and Google DeepMind, with results published on the Google Research blog. While the study shows promising capabilities, the authors note that responsible deployment will require additional validation and regulatory considerations. The work underscores the evolving role of AI in healthcare and the potential for advanced systems to augment clinical practice.