Google Introduces 6 Health AI Innovations, Starting From Nadi Detection To Cancer Research

Illustration of Google's AI Health solution (photo: Google)

JAKARTA - In The Check Up annual event which was held on March 18, 2025, Google announced the latest six artificial intelligence (AI) innovations aimed at helping people get more accurate health information.

First, with the updated AI Overviews technology, users can get more relevant and detailed answers related to various health conditions, including rare diseases. Google also introduced the "What People Suggest" feature, to help users discover experiences and suggestions from others who have similar conditions.

Second, to manage various information from various health sources, Google launched Medical Records API globally on Health Connect, which allows apps to read and write medical record information, such as allergies, medicines, immunizations, and laboratory results, in FHIR's standard format.

Furthermore, the Loss of Pulse Detection feature on Pixel Watch 3, will be launched in the US after previously available in 14 other countries. This feature can detect if the user has lost its pulse. So, if the user does not respond, this watch will automatically contact the emergency service.

Google also introduced AI Co-Scientist, a Gemini 2.0-based system that can help researchers analyze large amounts of scientific literature and produce new hypotheses. Interestingly, this tool has been tested by researchers at Imperial College London, Houston Methodist, and Stanford University.

"For example, if researchers want to understand the spread of disease-causing microbes, AI Co-Scientist will propose testable hypotheses, including a summary of relevant scientific literature as well as possible experimental approaches," Google wrote in its official blog.

Not only that, Google will also present TxGemma, Gemma-based open-model pool designed to help improve the efficiency of AI-based drug discovery, through the Health AI Developer Foundations.

TxGemma is claimed to be able to understand plain text as well as the structure of various therapeutic entities, such as small molecules, chemicals, and proteins, thus allowing researchers to ask questions to predict important properties of new therapeutic candidates, such as their level of security or effectiveness.

Lastly, in collaboration with the Princess Mryptxima Center in the Netherlands, Google helped develop Capricorn, an AI tool that accelerates the identification of personalized cancer treatment for children.

Using the Gemini model, Capricorn is capable of analyzing large amounts of medical data and produces a summary of treatment options that can help doctors make the best decisions for their patients.