Singapore Launches SIMFONI Medical AI Initiative for Locally Trained Healthcare Models
Source: The Straits Times
Singapore launched SIMFONI a national initiative to develop healthcare AI models trained on local patient data addressing the accuracy gap of Western-trained AI for Asian populations.
Singapore launched the Singapore Medical Foundation AI Model (SIMFONI) on July 9 a national initiative to develop healthcare AI systems trained specifically on local patient data and clinical practices addressing a critical gap in medical AI accuracy for Asian populations.
Health Minister Ong Ye Kung announced the programme at the NCS Impact 2026 conference noting that most AI foundation models used in healthcare today are trained on data from Western populations which limits their accuracy in Singapore clinical settings. The AI models will undergo rigorous evaluation tested on established medical benchmarks assessed against Singapore clinical guidelines and validated on local de-identified data from the public healthcare system. Executive director Professor Robert Morris said each candidate AI model goes through a rigorous evaluation and selection process tailored to Singapore context.
The programme will initially focus on two areas. First managing chronic diseases like diabetes high blood pressure and high cholesterol which represent the largest chronic disease burden on Singapore primary-care systems. Asians tend to develop diabetes at a lower body mass index compared with Westerners and genetic risk factors differ making locally trained AI essential. Second eye diseases such as cataracts retinal diseases and glaucoma where multimodal AI systems will process conversations eye images and medical records to support clinical decisions. The eye provides insights into broader conditions including diabetes and cardiovascular disease laying the foundation for expansion into other specialties.
SIMFONI is a programme under the Consortium for Clinical Research and Innovation Singapore (Cris) which brings together several MOH research programmes. Ong also outlined three prerequisites for deploying AI well in healthcare: a strong digital operating environment good quality data and a sound policy and organisational structure. MOH is in the final phases of replacing isolated IT systems across the entire public healthcare sector to enable integrated data sharing. The AI systems will offer possible diagnoses and treatment pathways but doctors will always make the final call.
Why it matters for Singapore: SIMFONI tackles a problem unique to Singapore and similar Asian healthcare systems where population-specific factors like lower-BMI diabetes risk and distinct genetic profiles make Western-trained AI models less reliable. By building AI that has effectively gone to Singapore medical school as Ong put it the initiative could set a template for how small high-income countries develop sovereign AI capabilities in healthcare. If successful it could reduce misdiagnosis rates improve chronic disease management and eventually expand into other specialties making Singapore a reference point for locally adapted medical AI globally.