Google's AMIE medical AI demonstrates real-time video consultation in first study

AMIE's advancement into real-time video consultation marks a shift in how AI can be applied to clinical work. Previous medical AI systems, including text-based versions of AMIE, operated within the constraints of written dialogue. The new multimodal architecture, built on Gemini and Project Astra, enables the system to process the full range of sensory information — visual and auditory — that human clinicians depend on when evaluating patients. The study's design, using patient actors and primary care physicians as evaluators, provides a controlled setting where AMIE's performance on core clinical competencies could be measured against human expectations.

The preference expressed by patient actors for video over text chat suggests that the addition of visual communication itself may carry therapeutic value beyond diagnostic accuracy alone. However, Google's explicit framing of AMIE as a research system underscores the distance between a favorable demonstration and clinical readiness. The company has not disclosed what additional research or validation steps remain, what regulatory pathways it may pursue, or whether AMIE will eventually be offered as a product or service.