SPEECH BY MR ONG YE KUNG, MINISTER FOR HEALTH AT THE SYNAPXE AI ACCELERATE CONFERENCE ON 16 JUNE 2025

Mr Tan Kiat How, Senior Minister of State for Health

Ms Jacqueline Poh, CEO, JTC Corporation

Board of Synapxe

Healthcare leaders, ladies and gentlemen

1.             A very good morning. It is my pleasure to join you at Synapxe’s inaugural AI Accelerate Conference today at Punggol Digital District, Singapore’s newest innovation hub.

Better Healthcare through AI

2.             As we gather here today, we find ourselves at a pivotal period in healthcare transformation. One major thrust in this transformation journey is to embrace healthcare technological advancement. There are several thrusts, but technology is driving many of the changes we are going to embark on. In particular, we are seeing the advent of precision medicine and the use of AI, which is today’s focus. 

3.             AI has become ubiquitous, surrounding us in everyday applications, and it has great potential to improve the delivery of healthcare. Our approach, which we must be very clear about, is to improve healthcare using AI, not implement AI for AI’s sake. 

4.             There must therefore be a sharp focus on use cases on practical applications of AI in transforming healthcare. We must deliver measurable outcomes, either in enhancing accuracy and speed of diagnosis, improving treatment or prevention, or increasing productivity for healthcare professionals. All our use cases have to pass these outcomes.

5.             We do not envisage AI replacing doctors and healthcare workers, but it should augment and enhance their capabilities. Our approach, therefore, is to ensure that care delivery is AI-enhanced or AI-enabled, but not AI-decided. The human must continue to be in the loop, making the final judgment on the most appropriate care pathway and medical intervention. I think patients also require that comfort that humans can give. Today, aeroplanes are very advanced and can fly by themselves. But you will be extremely worried if there is no pilot in the aircraft. Likewise, doctors, nurses and healthcare workers must be there.  In the process, we will also ensure continued patient safety, by implementing appropriate governance measures and safeguards.

Foundation Laid

6.             With this approach in mind, over the past year, we have laid a few foundation stones, which lay the groundwork for us to build our AI strategy. What are these foundation stones?

7.             First, is HEALIX (Health Empowerment through Advance Learning and Intelligent eXchange).  HEALIX enables secure sharing of up-to-date and anonymised clinical and socio-economic data, to build and train AI models for healthcare. We can think of HEALIX as the engine or factory to drive AI applications. The infrastructural platform is ready, and public healthcare institutions have been successfully onboarded.

8.             A second foundation stone is the HealthX Innovation Sandbox, set up by Synapxe. The Sandbox replicates a real-world environment using synthetic and anonymised data generated by HEALIX. This allows start-ups and innovators to experiment and develop practical use cases.

9.             The third foundation stone is to adopt the right regulatory approach. We always face disruptive technology. But if we continue to regulate sectors and services like we used to, sometimes disruptive technology will not see light. We need to adapt our regulatory approach. This is to accelerate innovation and adoption of AI, while maintaining robust clinical oversight and patient safety. To do so, we are setting up a a regulatory sandbox for AI software-as-medical devices (AI-SaMDs), where licensed public healthcare institutions can deploy the devices in-house without a manufacturer's licence or product registration. We are defining “in-house” very broadly, to deem all public healthcare institutions as a collective. They all come under the Ministry of Health (MOH), so all PHIs under MOH will be one system, and everything that is implemented within is considered in-house.

10.          This means an AI solution developed in one public hospital, MOH Office for Transformation or Synapxe can be adopted and quickly implemented in another public hospital. We will continue to ensure strong clinical governance, cybersecurity measures, and post-market surveillance post rollout of the AI solutions. If successful, we will explore also extending the sandbox to private healthcare.

Practical Use-Cases

11.          With these strong and necessary foundation stones, we are better able to develop and support the range of use cases. Many are ground-up, and used in our hospitals.  I think what MOH ought to focus on at this stage are the use cases that can be implemented throughout the system, such that it impacts every healthcare worker and patient. At this stage, it is important to focus on those impactful system-wide applications, because it will encourage acceptance by everybody, patients as well as healthcare workers. What are these? We actually have a clear idea. Let me share a few promising ones that are already in the pipeline. 

12.          Last year, I shared that we are using AI for chest X-rays in our hospitals, such as Changi General Hospital. Since then, we have been scaling AI to new chest X-ray use cases and care settings. The National Healthcare Group Geylang Polyclinic is using chest X-ray AI to help prioritise cases, so that patients with normal results can be managed more quickly and we focus more on those with more abnormal cases.

13.          This month, the National Centre for Infectious Diseases will start to use chest X-rays to screen for tuberculosis, and Woodlands Health will start to use imaging AI at its Emergency Department to detect bone fractures. 

14.          This capability will be rolled out progressively. It is quite a mature capability, but we need to work out the processes, ensure accessibility and ensure that people are trained. We want to roll this out across our public healthcare system. By end 2026, this should be a national capability. End 2026 is almost one and a half years away, and it is quite long. I think we should do it earlier, given how fast the technologies move. I am not looking at Synapxe, but at all our healthcare clusters.

15.          Next, in the realm of clinical operations, generative AI systems have become powerful tools to automate record-taking and documentation. Note Buddy, developed by SingHealth, has shown remarkable success. I tried it myself. It is, again, a very mature AI technology. It has already assisted over 2,100 healthcare workers in SingHealth to create more than 16,000 medical and administrative notes. This has raised productivity and reduced toil. All public healthcare institutions will be using this or similar tools by the end of this year.

16.          There is also great potential to incorporate AI capabilities into patient-facing health apps. The most commonly used health app today is HealthHub. We are working to consolidate HealthHub with our cluster-specific apps – Health Buddy, NHG Cares App, and NUHS App. By 2027, there will be one unified platform to improve online accessibility of public healthcare services.

17.          We can enhance this newly consolidated HealthHub with AI, so that the app can do more than manage health records, appointments, and bill payments. HealthHub AI can offer personalised health information, such as how to manage common health conditions, including through lifestyle changes. This is conveyed through speech and text, in four languages and in a conversational style. I believe using agentic AI can also help you book appointments, with you just telling the App to do so. It is currently in beta and users’ feedback has been promising. Synapxe aims to launch this later this year.

18.          Finally, predictive preventive care, the next frontier in healthcare transformation. At some point in the future, AI models combined with the judicious use of medical, socio-economic, and even genetic data can help doctors predict if someone is likely to develop serious conditions many years in advance.

19.          This will be in the next bound of Healthier SG, leveraging health records, evidence-based clinical risk scoring and AI's predictive powers to support preventive care and social prescribing. While we harness the power of AI, we will continue to maintain the human touch through our family doctors, guiding and empowering residents to take charge of health outcomes before illness strikes. We are not far away from this future. We may not predict every single disease. The major diseases in Singapore affecting us and the healthcare system the most are stroke, cardiac arrest, common cancers, the “three Highs” and kidney failure. If we are able to sieve out who are at highest risk of getting these most common serious diseases, we can take a huge leap in terms of preventive care and improve the Healthier SG system as we know it.

Closing

20.          AI heralds the beginning of a new technological era. Its ability to plough through human records, observations and knowledge in split seconds to provide an insight or conclusion makes it a very powerful entity in economy and society. 

21.          Yet AI is not human. It does not feel, think, hope, fear or empathise. If you ask AI “Do you feel?”, it replies you “Obviously I don’t, I just crunch data”. Hence, as much as airplanes can fly themselves now, we will not accept a plane without a pilot. Likewise, while there is literature suggesting that AI provides better medical prognosis and diagnosis and treatment plan than a doctor equipped with AI, I think most of us will still prefer the latter because we trust the human doctor who can empathise with our worries and concerns.

22.          AI therefore expands but can also subtract. It expands our skills and ability to get more things done and better. It also subtracts the human experience of discovery, trial and error, making mistakes and learning from a fellow human being. 

23.          How we fully use and leverage AI, but not become subservient or mindlessly defer to it, is a balance we must strike. This balance is most apparent in healthcare, I cannot think of a sector where the trade-off is more stark than in healthcare. Because this is a realm where knowhow is driven by science and evidence, but every act is delivered with human kindness and empathy.

24.          Whether by coincidence or divine intervention, AI is bursting into the scene just when human societies are experiencing an unprecedented process of ageing. Patient numbers are rising, with more complex and multiple conditions; healthcare systems all over the world are stressed; care is shifting to the community and patients’ homes; prevention of diseases is taking centre stage. AI at this point is a scientific gift to us, to tackle the biggest social challenge of our time. 

25.          I thank Synapxe for playing a part in driving AI in healthcare. However, this is not purely a technological endeavour. Let us remain committed, considered, yet bold in harnessing AI to uplift patient care, empower healthcare professionals, uphold the ethos and essence of healthcare, and shape a future where everyone can live healthier, more fulfilling lives. Thank you, and all the best in this exciting journey.