Singapore's AI Governance Framework: A Blueprint for Healthcare Innovation
How Singapore's Model AI Governance Framework for Agentic AI positions the city-state as a global leader in responsible healthcare AI deployment.
A Landmark Framework for the Agentic AI Era
On 22 January 2026, Singapore's Minister for Digital Development and Information, Mrs Josephine Teo, announced the launch of the Model AI Governance Framework for Agentic AI at the World Economic Forum in Davos. Issued by the Infocomm Media Development Authority (IMDA), it is presented as a first-of-its-kind framework addressing AI agents that can independently plan, reason, and take autonomous actions on a user's behalf. It builds on the Model AI Governance Framework for AI that IMDA first introduced in 2020.
The framework is voluntary. It is organised around four dimensions: assessing and bounding the risks upfront, ensuring meaningful human accountability, implementing technical controls and processes, and enabling end-user responsibility. IMDA emphasises that compliance is not mandatory but that organisations remain legally accountable for their agents' behaviour and actions. The guidance applies to agentic AI deployments across sectors, including healthcare.
The framework arrives as Singapore continues to concentrate AI activity in Southeast Asia. According to DealStreetAsia, Singapore attracted about $4.6 billion in total venture funding in 2025, of which roughly $1.4 billion — close to a third — went to AI-related startups, even as overall venture funding contracted year-on-year. This governance work also extends the direction set by Singapore's second National AI Strategy (NAIS 2.0), launched in December 2023, which framed AI governance as a national priority alongside AI adoption.
Pensieve-AI: A Case Study in Governed Healthcare Innovation
The pairing of innovation support with governance is exemplified by Pensieve-AI, a locally developed tool for detecting pre-dementia. It was developed by Singapore General Hospital, working with the Government Technology Agency of Singapore (GovTech) and collaborators at Duke-NUS Medical School and the National University of Singapore. The test is self-administered on a touchscreen tablet in under five minutes and comprises four drawing tasks, including a clock-drawing task, which the app's AI then analyses for signs of cognitive impairment.
In a study of 1,758 community-dwelling adults aged 65 and older, published in Nature Communications in March 2025, Pensieve-AI achieved an AUC of 0.93 (93.1%) in detecting mild cognitive impairment and dementia — performance the authors describe as comparable to detailed neuropsychological testing. The tool is slated for wider rollout in Singapore in 2026.
Pensieve-AI illustrates why governance and clinical ambition are often framed in Singapore as complementary rather than competing goals: a tool that seniors can complete themselves in minutes only earns adoption if patients and clinicians trust how it handles data and how its outputs are acted upon. That trust is exactly what the country's governance instruments are designed to underwrite.
The Regulatory Stack Behind Healthcare AI
Singapore's agentic AI framework sits on top of an existing set of instruments that already govern how AI systems handle personal data and operate in clinical settings. Together they form the practical compliance backdrop for any healthcare AI deployment in the city-state.
On data, the Personal Data Protection Commission (PDPC) published its Advisory Guidelines on the Use of Personal Data in AI Recommendation and Decision Systems on 1 March 2024. The guidelines are advisory rather than legally binding, but the PDPC has signalled it will enforce the Personal Data Protection Act consistently with them. They clarify when organisations may rely on exceptions such as the Business Improvement and Research exceptions, and set out data-handling and accountability expectations across development, deployment, and procurement. Notably, they focus on recommendation and decision systems and do not address generative AI, and they are not healthcare-specific.
On clinical use, the Ministry of Health and the Health Sciences Authority maintain the Artificial Intelligence in Healthcare Guidelines (AIHGle), first issued in 2021 and updated to version 2.0 in March 2026. AIHGle sets out core ethical principles for healthcare AI — including safety, fairness, transparency, explainability, robustness, security and data protection — and clarifies the responsibilities of developers, deployers (healthcare organisations), and users (healthcare professionals). It is advisory and complements the HSA's regulatory requirements for software as a medical device. For a tool like Pensieve-AI, these instruments — PDPA obligations, PDPC guidance, and AIHGle — define the governance envelope within which deployment happens.
Implications for the Global Healthcare AI Landscape
Singapore's agentic AI framework is notable less for being binding — it is not — than for being specific to agentic systems at a moment when most jurisdictions' AI rules were drafted before autonomous, multi-step agents reached their current capabilities. By concentrating on how to bound risk and preserve human accountability rather than on blanket approvals or bans, it offers one reference point among several emerging approaches to agentic AI governance.
Its four dimensions — bounding risk upfront, meaningful human accountability, technical controls and processes, and end-user responsibility — map onto questions every healthcare deployer eventually faces, regardless of jurisdiction: what the agent is allowed to do, who is answerable when it errs, what technical guardrails constrain it, and what the end user is expected to verify.
For organisations entering the Asian healthcare AI market, the practical takeaway is that Singapore has published unusually concrete expectations across the stack — data protection, healthcare-specific ethics, and now agentic AI — even where those expectations remain voluntary. Building to them early is less about satisfying a single regulator today than about being positioned for how healthcare AI governance is likely to tighten across markets.
Sources
- Infocomm Media Development Authority (IMDA), "Singapore Launches New Model AI Governance Framework for Agentic AI," 22 January 2026, https://www.imda.gov.sg/resources/press-releases-factsheets-and-speeches/press-releases/2026/new-model-ai-governance-framework-for-agentic-ai
- Ministry of Digital Development and Information (MDDI), "Singapore Launches New Model AI Governance Framework for Agentic AI," 22 January 2026, https://www.mddi.gov.sg/newsroom/singapore-launches-new-model-ai-governance-framework-for-agentic-ai--/
- Liew TM et al., "PENSIEVE-AI: a brief cognitive test to detect cognitive impairment across diverse literacy," Nature Communications, March 2025, https://www.nature.com/articles/s41467-025-58201-x
- DealStreetAsia, "AI investments made up a third of Singapore venture funding in 2025," 2026, https://www.dealstreetasia.com/stories/singapore-venture-funding-2025-483749
- Ministry of Digital Development and Information (MDDI), "National Artificial Intelligence Strategy 2.0," 4 December 2023, https://www.mddi.gov.sg/newsroom/04122023/
- Personal Data Protection Commission (PDPC), "Advisory Guidelines on Use of Personal Data in AI Recommendation and Decision Systems," 1 March 2024, https://www.pdpc.gov.sg/news-and-events/announcements/2024/03/advisory-guidelines-on-use-of-personal-data-in-ai-recommendation-and-decision-systems
- Ministry of Health (MOH) & Health Sciences Authority (HSA), "Artificial Intelligence in Healthcare Guidelines (AIHGle)," first issued 2021, v2.0 March 2026, https://www.moh.gov.sg/others/health-regulation/emerging-regulatory-policy-issues/
Related Articles
ASEAN AI Adoption: From AI-First to AI-Native in 2026
Southeast Asia's AI landscape is maturing rapidly, with Singapore leading investment, Vietnam and Indonesia building sovereign AI capabilities, and the region confronting persistent skills and infrastructure challenges.
Source: Google, Temasek, and Bain & Company, "e-Conomy SEA 2025 Report"
Market AnalysisAI Healthcare Innovation Across Southeast Asia: Thailand, Vietnam, and Indonesia
Beyond Singapore, Southeast Asia's largest economies are charting distinct paths to AI-powered healthcare, shaped by their unique demographics, infrastructure, and strategic priorities.
Source: Thailand National AI Strategy and Action Plan (2022-2027), NECTEC/NSTDA and Ministry of Digital Economy and Society (MDES)
ASEANSingapore's Smart Nation Initiative: How AI Agents Are Transforming Public Healthcare
Singapore is translating national digital policy into practical healthcare outcomes by deploying AI agents across hospitals, polyclinics, and community care programs with governance built in.
Source: Synapxe, "New Singapore Standards to Strengthen the Interoperability of Digital Solutions for Public Healthcare" (SS 719:2025 and SS 720:2025), 31 October 2025 — https://www.synapxe.sg/media-releases/innovation/new-singapore-standards
Building with Agentic AI?
Learn how Ajentik's autonomous agent platform is helping enterprises deploy production-ready AI agents at scale.
Schedule a Consultation