Google has demonstrated an AI system holding expert-level clinical video consultations for the first time. In a randomized study published by Google Research on August 12, AMIE (Video) — the video configuration of the Articulate Medical Intelligence Explorer, built on Gemini and Project Astra — was rated by an independent panel of 20 primary care physicians as on par with human doctors across core clinical competencies.

The study covered 100 clinical scenarios across five body systems, with 15 trained patient actors carrying out 300 standardized consultations. AMIE (Video) matched 10 board-certified physicians in history-taking thoroughness, diagnostic accuracy, management appropriateness and communication quality — and was rated significantly higher than both the doctors and the text-only AMIE at eliciting physical signs and guiding patients through virtual examination maneuvers.

Architecturally, AMIE (Video) splits the workload across three specialized agents running in parallel: a patient-facing Talker that keeps conversational latency natural, a background Planner that continuously refines the differential diagnosis, and a Perception agent that monitors the audio and visual streams for clinically relevant non-verbal cues such as visible distress or physical findings.

Patient actors strongly preferred the video interface over text chat, rating it as significantly easier to use and more effective for communicating health concerns, and scored AMIE (Video) favorably on empathy, rapport and confidence in care. Google stresses the limits: the study used professional actors in simulated settings, not real patients, and intermittent technical issues remain. Real-world validation is already underway, including a feasibility study with Beth Israel Deaconess Medical Center and an ongoing nationwide randomized study with Included Health.

Sources