Google has unveiled AMIE Video, an advanced version of its medical artificial intelligence system designed to conduct clinical consultations over live video.

The research system, powered by Google’s Gemini models and Project Astra, can simultaneously analyze a patient’s voice, video and visual behavior during a conversation. It is designed to identify potential signs such as distress, changes in breathing or difficulty moving.

Unlike text-based medical AI systems, AMIE Video is designed to more closely simulate a remote doctor visit, allowing it to hold a natural conversation with a patient while also guiding the patient through virtual physical examinations.

The system uses several AI agents working simultaneously. One manages the conversation, another develops an overall picture of the patient’s medical situation, while another analyzes video and audio information in real time.

In a controlled Google study involving 100 clinical scenarios and 300 consultations, physicians evaluated AMIE Video and found that it achieved expert-level performance across several measures.

Google emphasized, however, that AMIE Video remains a research system and is not currently intended for use in actual patient care.

Testing so far has been conducted in simulations involving actors portraying patients rather than in real-world clinical settings.

Before such technology could be deployed for medical care, it would require additional research and likely regulatory approval as a medical device.

For now, AMIE Video offers a glimpse at how artificial intelligence could eventually transform telemedicine by allowing AI systems to analyze not only what patients say, but also what can be seen and heard during a medical consultation.

(YWN World Headquarters – NYC)