Singapore's CGH: AI Literacy, Not Just Governance Frameworks, Is Key to Safe Healthcare AI

Singapore's public hospitals are running hundreds of AI projects, but the hardest problem isn't building the models — it's making sure the people using them actually understand what they can and can't do. Changi General Hospital's Deputy Director of Data Science & Intelligence, Narayan Venkataraman, made this case at HealthTechX Asia 2026, arguing that AI governance frameworks alone are not enough. Without organisation-wide AI literacy, he warned, transformation efforts will fail.

Speaking at the conference, Venkataraman shared CGH's approach to AI governance, which goes beyond the usual checklists of risk assessment and compliance. The hospital's strategy is grounded in the principle that everyone has a role to play in managing AI risk — from clinicians and nurses to pharmacists, operational leaders, and executives. Responsibility for AI governance, he argued, does not sit solely with technical teams. This is a deliberately inclusive model, and it reflects a reality that many Singapore organisations are only beginning to confront: AI is no longer a tool used by data scientists alone, but one that frontline staff interact with daily.

To put this into practice, CGH conducts regular organisational pulse checks to assess AI fluency levels, usage behaviours, and knowledge gaps across the hospital. Tailored training sessions are designed to bridge those gaps — including hackathons that encourage experimentation with AI tools and hands-on workshops that examine real-world failures rather than success stories. One telling example Venkataraman shared: an AI-generated image of Singapore's skyline showed Marina Bay Sands with four towers instead of three. The image looked convincing, but it was wrong. The point was simple but important — AI can appear correct while still being incorrect, and staff need the literacy to recognise when to question its outputs.

This matters because CGH, like the rest of Singapore's public healthcare cluster, is moving from the experimentation stage of AI maturity toward deeper integration into clinical workflows. The migration of Singapore's public hospitals to a national electronic medical record platform will accelerate this shift, embedding AI decision-support tools directly into how care is delivered. Venkataraman drew on publicly available guidance from IMDA — including the Model AI Governance Framework and Generative AI Governance guidelines — as reference points that organisations can adapt to their own environments.

Why it matters for Singapore: As Singapore's healthcare system digitalises rapidly, the gap between AI capability and AI literacy is becoming one of the most significant risks to safe adoption. CGH's model — combining data governance, AI literacy programmes, and shared accountability across all staff levels — offers a blueprint that other Singapore institutions can follow. The message from HealthTechX 2026 is clear: having the best AI tools means little if the people using them cannot judge when the tools are wrong. AI literacy needs to be treated not as a nice-to-have add-on to governance, but as its foundation.