Can chatbots serve doctors and patients?

Healthcare can involve navigating complex journeys, and this is a challenge we hope to help solve using artificial intelligence (AI). This notebook explores Google’s research at the cutting-edge intersection of AI and medicine, drawing on advancements in models like Google’s AMIE and MedGemma. We explore how the AMIE system is being developed to assist with clinical workflows spanning both general and subspecialist domains, from single-visit to multi-visit cases, and analyze multimodal data including images and documents. We also dive into how Personal Health Large Language Models (PH-LLM) can help interpret wearable sensor data to provide personalized health insights and answer objective health queries, and how developers can leverage a growing set of open models within the HAI-DEF family, such as MedGemma, to develop AI for their applications.