Singapore healthcare AI push moves into clinical and regional deployment

Singapore’s ageing population is pushing healthcare AI beyond research pilots and into a practical question for the system: whether digital tools can help clinicians detect risks earlier, support diagnosis, and extend care capacity without weakening patient safety.

Minister of State for Digital Development and Information & Health Mdm Rahayu Mahzam outlined that direction at AI in Health x ATxSummit on 19 May, where two MOUs were announced across clinical AI development, research translation, and regional healthcare collaboration with Bhutan.

By 2030, Singapore will have more seniors than children for the first time, with one in four Singaporeans aged 65 or older. The government is tying AI adoption to a wider move upstream in healthcare, using data to identify health patterns earlier and building on Healthier SG’s focus on prevention and community care.

Ageing pressure sets the agenda

Singapore’s healthcare AI work is being shaped by two pressures at once: a rapidly ageing population and the need to deploy AI in ways that fit clinical realities.

Mdm Rahayu said Singapore has built digital foundations across healthcare IT systems, data infrastructure, and institutional trust. Those foundations give the country a base to apply AI to prevention, diagnosis, and care delivery, rather than treating the technology as a standalone experiment.

The “Smart Nation to Blue Zone Nation” theme placed AI within a broader health outcome agenda. The reference to Blue Zones, where communities often live well beyond 90, points to a model of healthcare that depends on prevention and daily health behaviours as much as clinical intervention.

For Singapore, the operational challenge is translating that ambition into tools that clinicians can use and patients can benefit from.

Research moves closer to clinical use

The first MOU, between Singapore General Hospital and A*STAR’s Diagnostics Development Hub, focuses on bringing research closer to patient impact. The partnership has supported three innovations: the in-vitro Antibiotic Combination Test, PENSIEVE-AI, and HealthVector Diabetes.

The in-vitro Antibiotic Combination Test helps doctors select antibiotic combinations for patients while addressing antimicrobial resistance. PENSIEVE-AI is a digital drawing application that identifies early memory problems in seniors.

HealthVector Diabetes is the most directly tied to Singapore’s public health burden. Mdm Rahayu described it as the world’s first digital twin model of human biology that can estimate the risk of chronic kidney disease in people with type 2 diabetes.

Diabetes remains one of Singapore’s most pressing health challenges, with complications including kidney failure, blindness, and amputation. A tool that can identify high-risk individuals earlier would be relevant only if it can support timely intervention within actual care workflows.

Bhutan partnership tests regional application

The second MOU, between SingHealth and Gyalpozhing College of Information Technology, Royal University of Bhutan, extends Singapore’s healthcare AI work into a regional setting.

The collaboration will develop an AI-assisted chest radiograph model trained specifically on Bhutanese data. It is intended to support the diagnosis of lung diseases, including infections and cancer, in rural hospitals.

The model is built on MerMED-FM, which was co-developed by SingHealth and A*STAR. The use of Bhutanese data is significant because healthcare AI models often need to account for local clinical patterns, healthcare settings, and population characteristics before they can be applied safely.

The partnership also covers guidelines, educational programmes, and regulatory frameworks for responsible AI use in healthcare, tailored to Bhutan’s context. That gives the collaboration a governance and capability-building layer, rather than limiting it to model deployment.

Governance stays close to deployment

Clinical AI adoption depends on more than model performance. Mdm Rahayu linked deployment to trust, growth, and community, and pointed to governance as a condition for patient safety.

SingHealth has developed the S.C.O.R.E. framework to evaluate large language model outputs and AI-generated responses in clinical settings. The framework covers Safety, Context & Consensus, Objectivity, Reproducibility, and Explainability.

Clinical teams across SingHealth have applied S.C.O.R.E. to evaluate AI tools that support patient care delivery, including medication enquiry chatbots and AI systems that assist specialists during consultations. It has also been used in clinical decision support to validate model outputs before deployment and guide the selection of suitable models for different clinical contexts.

The Singapore Government has also contextualised guidance such as the AI in Healthcare Guidelines and Model AI Governance Frameworks to guide safe and responsible AI deployment in healthcare.

Healthcare is one of Singapore’s national AI Missions. The next stage will depend on how well the country connects research, clinical workflows, governance, and regional deployment into systems that clinicians can trust and patients can use.