Taiwan is building something that could change how countries handle public health. The island’s health system now uses Google’s AI to catch diabetes risk early, turning a 20-minute doctor assessment into a 25-second computer analysis.
The partnership between Google and Taiwan’s National Health Insurance Administration takes 20 years of patient data and turns it into early warnings. Soon, 10 million Taiwanese will have an AI health assistant on their phones that gives personalized medical advice based on government guidelines.
How will it work?
The system centers on something called AI-on-DM (Artificial Intelligence on Diabetes Mellitus). Before this tool existed, screening one patient for diabetes risk took a doctor about 20 minutes. Checking 20,000 people meant 40 health professionals working non-stop for three weeks.
Now the AI model does the same work in 25 seconds per patient. The entire 20,000-person screening takes under 90 minutes. That’s a 14,400-fold speed increase.
The system works by:
- Analyzing patient data from Taiwan’s unified health database
- Flagging potential risk patterns for doctors to review
- Helping doctors intervene before serious complications develop
- Processing everything through Google Cloud’s computing power
This month, Taiwan launches the second part: a Gemini-powered health assistant. The AI tool will work inside the government health app that 10 million Taiwanese already use. It gives personalized health suggestions based on each person’s medical history and official clinical guidelines.
Why does it matter?
Even Taiwan’s excellent health system has limits. Doctors only have so many hours each day, and manual risk assessments eat up time that could go toward treating patients who need immediate help.
The AI system ensures every citizen gets the same quality health screening, whether they live in Taipei or a rural village. It also frees up medical experts to focus on high-risk cases and complex patient care that actually needs human judgment.
For diabetes specifically, early detection makes a huge difference. The disease causes serious complications when left untreated, but catching it early allows for much better outcomes.
The context
This builds on several AI health projects already running in Taiwan’s hospitals:
- China Medical University Hospital uses MedLM for cancer care
- Chang Gung Memorial Hospital has AI-enhanced ultrasound diagnostics
- Taipei Medical University’s hospital uses automated workflows to handle doctor shortages
- The health administration processed over 30,000 pathology reports using MedGemma
Google.org also gave $1 million to Taiwan’s Digital Humanitarian Association. The money will bring diabetes management services and digital training to 300 community centers. The program aims to support 240,000 health check-ins and train 200 local diabetes caregivers.
Taiwan plans to expand this same AI framework to screen for high blood pressure and cholesterol problems. The country’s single-payer system and comprehensive patient database make it an ideal testing ground for population-scale AI health tools that other countries could copy.