🤖 Google AMIE (Video): AI doctor for video consultations at the level of real specialists
Google Research and DeepMind have introduced AMIE (Video) based on Gemini. Three-agent architecture: Talker Agent conducts the dialogue, Planner Agent updates the diagnosis, Perception Agent analyzes video and audio. In an OSCE study of 300 consultations, the system is comparable to 10 primary care doctors. Actor-patients preferred the video mode to chat. Paper: arXiv 2608.09861.
🌍 Asynchronous three-agent architecture — a reference model for separating low-latency dialogue and clinical reasoning.
👤 Demonstration of the transition of multimodal AI to video consultations. The system is research-oriented.
Source: https://research.google/blog/advancing-amie-towards-expert-level-audio-visual-clinical-consultations/ Paper: https://arxiv.org/abs/2608.09861
