Google is turning Singapore into a test bed for medical AI. At the first Google for Health Summit, held in the city-state on October 29, the company said it will open a Google Research hub there and announced a set of projects with two of Singapore's public healthcare groups, covering clinical decision support, preventive care and disease surveillance.
The hub will be staffed by researchers and software engineers based in Singapore. Google says the team will work on foundational AI models aimed at speeding up scientific discovery, and Yossi Matias, who runs Google Research, told the audience it will collaborate with the company's global teams with a focus on making frontier models more efficient, according to The Business Times. The group is also expected to partner with Nanyang Technological University and the National University of Singapore on talent and research funding, and to carry existing health work such as the open MedGemma models into frontline care, with Google stressing that data would stay under local control.
The most clinically sensitive piece involves Google DeepMind and the National University Health System (NUHS). The two will evaluate DeepMind's AI co-clinician, a research tool that pulls together medical guidelines and published literature to answer doctors' questions while they treat patients. The Business Times reports that the work is meant to measure how accurate, safe and useful the system's answers are against established clinical criteria. Google describes this as an evaluation, not a deployment.
A second project pairs Google Cloud with NHG Health on population health. Predictive models are meant to flag groups of residents at higher risk of chronic illness, and those insights then feed an AI health coach inside the NHG HealthBot, which runs on Google's Gemini Enterprise for Customer Experience platform and sends lifestyle advice and preventive care reminders. According to The Business Times, the model combines NHG's de-identified population records with Google's geospatial tools and may draw on climate, environmental, search and behavioural signals to anticipate conditions such as frailty. Neither party has published how those signals will be governed or how residents can opt out.
There is also money for public health surveillance. Google.org and Google DeepMind are giving $7 million to the Philanthropy Asia Alliance's Health for Human Potential Coalition, of which $1 million goes to PathGen, a platform that helps countries spot disease threats sooner and pool public health data without moving it across borders.
The announcements build on a national AI partnership between Google and Singapore that was announced earlier this year, aimed at applying Google's frontier models to a range of public challenges. Ben King, who heads Google in Singapore, said the company wants to build health AI locally and later take it to other markets.
Singapore's government set out conditions on the same stage. Tan Kiat How, a Senior Minister of State whose portfolio spans health as well as digital development and information, said clinical AI must pass rigorous testing before deployment and be monitored afterwards, and warned that a capable model does not by itself make people healthier. No timelines, patient numbers or outcome targets were disclosed for any of the projects, so the summit marks a set of commitments and trials rather than proven results.