Google.org hosted a workshop that united 50+ organizations to connect, collaborate, and advance the future of Health AI. Over three days, the workshop moved from keynotes and product demos to benchmarking exercises and discussions on public health, drug discovery, AI for disease screening, and more.
About us
Google Health is committed to helping everyone live more life every day through products and services that connect and bring meaning to health information. We’re developing technology solutions to enable care teams to deliver better, faster and more connected care. We’re working on products and features to empower people to be healthier with the information, assistance, and connections they need to act on their health. And we’re exploring the use of artificial intelligence to assist in diagnosing cancer, predicting patient outcomes, preventing blindness and much more. Our work complements Google’s mission to organize the world’s information and make it universally accessible and useful
- Website
-
https://health.google/
External link for Google for Health
- Industry
- Technology, Information and Internet
- Company size
- 10,001+ employees
Updates
-
The global health workforce is projected to face a shortage of 11 million workers by 2030. We’re exploring how Google's AI models could help address this challenge by serving as helpful tools in medical learning environments. In feasibility studies, medical learners and educators evaluated LearnLM, Google's set of Al models infusing Gemini with learning science principles. Across 50 medical education scenarios, LearnLM was preferred over the base model, with physician educators judging it “more like a very good human tutor.” With these capabilities now integrated into Gemini 2.5 Pro, we see potential to accelerate clinical competency and reimagine health professions education with Al. Learn more from Google Research: https://goo.gle/3JAQvqD
-
Introducing a new Al-powered personal health coach— a fitness trainer, sleep coach, and health advisor all working together to help you be your best. Backed by research and built with industry experts, the coach is designed to deliver guidance that’s uniquely tailored to your health goals and real-life circumstances. It was developed with the invaluable guidance of our Consumer Health Advisory Panel, a diverse group of leading experts spanning medicine, artificial intelligence, behavioral science, and more. Their scientific and clinical insights ensure our features are grounded in evidence and designed to genuinely improve your well-being. Starting this October, we'll begin rolling out a preview of the personal health coach as part of Fitbit Premium in the redesigned Fitbit app. Learn more: http://g.co/health/phc
-
From smartwatches to diagnostic tests, health data is increasingly multimodal— often carrying unique and overlapping signals. To analyze them together, Google Research developed M-REGLE, an AI method that allows for the joint analysis of multiple types of label-free health data. This work contributes to uncovering genetic links to disease and improving our ability to predict risk. https://goo.gle/45nkw4t
-
-
Google's research on a Personal Health Large Language Model (PH-LLM) is now published in Nature Medicine Magazine! Mobile and wearable devices offer continuous personal health data, such as step counts, heart rate variability and more. Gen AI can enhance the use of this data to support personal health and wellness tracking, but needs to be able to interpret complex data and apply relevant health domain knowledge to produce context-aware insights and recommendations. In this new publication, Google Research engineers evaluated PH-LLM on multiple sleep and fitness tasks, demonstrating that the model could exceed the average scores of human experts on professional exams, effectively analyze real-world coaching case studies, and predict self-reported sleep quality from wearable sensor data. While further development and evaluation are necessary, this research demonstrates both the broad knowledge and capabilities of Gemini models for personal health applications. Learn more: https://goo.gle/3Hvhwen
-
-
🚨 New research from Google Research and Google DeepMind: Towards physician-centered oversight of conversational diagnostic AI. Conversational AI holds promise for improving diagnostic workflows, but safe, real-world use must ensure that licensed physicians stay accountable for clinical decisions. In this new publication, Google introduces a framework for asynchronous physician oversight of AMIE. Guardrailed-AMIE, or g-AMIE, performs history taking but abstains from offering medical advice and conveys assessments for review through a “clinician cockpit” to an overseeing primary care physician, who maintains full oversight and accountability. In a randomized, blinded virtual OSCE study across 60 cases, primary care physicians (PCPs) overseeing g-AMIE produced higher-quality composite decisions than nurse practitioners, physician assistants, or even other, less experienced PCPs, working under the same constraints. Read more: https://goo.gle/4lo616o
-
In health research and clinical care, critical information is often buried in unstructured text—from detailed clinical notes to the latest research papers. Manually processing this data is not only a major time commitment, it can also lead to missed insights. To help our partners in the research and developer communities tackle this challenge, we’re introducing LangExtract. It's a new, open-source Python library designed to turn large volumes of unstructured text into structured, actionable data. To see these capabilities in action, we've built an interactive demo, illustrating how LangExtract can parse the findings section of a radiology report into schema-controlled data. Learn more about LangExtract: https://goo.gle/3IOSerY Try the interactive demo for yourself: https://lnkd.in/gCt9wagv
-
Since introducing MedGemma, we’ve been excited to see how developers might use our open-weight model to build helpful healthcare applications. To showcase MedGemma's capabilities, we built a demo that makes use of MedGemma for medical education. The demo showcases how MedGemma can guide a medical student through chest x-ray interpretation, combining advanced reasoning and multimodal capabilities. The model presents targeted multiple-choice questions, integrates relevant clinical guidelines via Retrieval-Augmented Generation (RAG), and provides clear rationales for its interpretation, deepening the student’s understanding through comparative analysis. This demo is one example of how open-weight models like MedGemma can be used to create sophisticated, personalized tools that support clinicians and researchers. See the potential in action: https://goo.gle/4mhWVci
-
-
In just 12 hours, 130 experts at Google France’s health hackathon developed 26 prototypes exploring how open AI models can potentially help tackle healthcare’s biggest challenges. From oncology support to ER triage and beyond, the projects show how collaboration and open-weight AI models like MedGemma and TxGemma can help create tangible improvements in healthcare. Check out the winning projects from the hackathon → https://goo.gle/44YYfJY
-
Meet Google's Consumer Health Advisory Panel. We've convened experts across medicine, digital health, AI, wearables, physical activity, sleep, and behavior change to help us uphold the highest standards for the research and development of health features. We're committed to evidence-based research and product development, especially for novel health technologies, and the Consumer Health Advisory Panel is an essential part of our approach. Learn more: https://goo.gle/4lveGEX