Finding health risks earlier can make all the difference. That’s the goal of a collaboration between Google and Taiwan’s National Health Insurance Administration (NHIA), which uses Gemini to help doctors analyze data from Taiwan’s overall health system. Even with one of the best healthcare systems in the world, doctors’ time is limited, and artificial intelligence can help them make the most of it. Today we’re sharing how Gemini technology is helping doctors detect health risks earlier than ever before. Taiwan’s health system is a global leader with a single payer and population health database that has helped care for nearly all Taiwanese for over 20 years. But even with big data, doctors only have so many hours in the day.
From weeks of work to a few seconds
The first major milestone in this collaboration is AI-on-DM (Artificial Intelligence on Diabetes Mellitus), an AI model by NHIA under the Ministry of Health and Welfare (MOHW) that helps assess diabetes risk. Previously, a risk assessment for a single patient took an average of 20 minutes. Screening 20,000 people would have required 40 professionals three weeks of non-stop work. By using the AI-on-DM model, which digitizes clinical logic and leverages Google Cloud’s concurrency, processing time is reduced to just 25 seconds per case. case. This 14,400x efficiency boost enables the system to complete 20,000 evaluations in under 90 minutes, making population-scale risk assessment a scalable public health tool.
More time for what matters
The model flags potential data patterns for clinical review and helps doctors to help them intervene before complications arise. Soon, patients will also have expert support in their pocket. This month, NHIA will launch a Gemini-powered health assistant that will generate personalized insights on its government app used by 10 million people in Taiwan. The tool provides personalized, safe suggestions based on clinical guidelines to help people with their daily care.
This shift ensures that expertise is focused where it matters most: high-risk interventions and people-centred support. It also ensures that all citizens receive the same assessment of high quality, regardless of where they live.
The AI-on-DM project builds on successful AI collaborations with hospitals across Taiwan. Key milestones include CMUH’s adoption of MedLM for cancer care, Chang Gung Memorial’s AI-enhanced ultrasound diagnostics, and TMU-affiliated hospital’s use of automated workflows to alleviate clinic shortages. NHIA also used MedGemma to process over 30,000 pathology reports.