Securing Singapore’s Future.
AI can help Singapore sharpen its competitive edge — strengthening its position as Southeast Asia’s hub for trade and innovation. It can accelerate breakthroughs in medical research, and help the country navigate the long-term fiscal and workforce pressures of an ageing population.
AI will support Singapore’s position as Southeast Asia’s regional trade hub.
Since its founding, Singapore has thrived as a hub for regional and world trade. Today, Singapore remains one of the world’s most open economies, while the overall performance of its economy has been strongly correlated to that of the wider Southeast Asia region in recent decades.
Our research identified three core mechanics by which AI and trade could continue to benefit the wider Singaporean economy:
Help companies navigate non-tariff barriers to trade.
Businesses can find it challenging to navigate the regulatory requirements of exporting beyond their borders. Even the complications of filling in paperwork in different languages becomes a serious point of friction. Conversely, AI is typically well suited for supporting this kind of administrative task. By helping ease current barriers to trade from language or regulatory submissions, we estimate AI could help boost exports for Singapore by over S$380 billion (US$283 billion).
Rising Southeast Asia tide for regional trade.
AI is likely to lead to leapfrog growth in many of the less developed economies in Southeast Asia – helping accelerate their wider economic development. This in turn is likely to increase their economic demand and inter-regional trade as a whole, with Singapore as the central hub of this. Over the next ten years, AI is likely to increase the GDP of the whole Southeast Asia region by almost S$360 billion (US$270 billion).
Enhanced supply chain optimisation.
AI-driven logistics could help optimise shipping routes, warehousing, and inventory management. This in turn could lower costs and enhance reliability and help offset supply chain disruption. Early surveys have found that some of the highest cost reductions from AI are coming in supply chain management11, while early adopters of autonomous supply chain planning report12 that it can “lead to an increase in revenue of up to 4 percent, a reduction in inventory of up to 20 percent, and a decrease in supply chain costs of up to 10 percent.”
Map displays value of exports to Southeast Asia supported by AI.
SEA-LION
is helping to improve the accuracy and experience of using AI across Southeast Asia.
Skilled Singaporean workers can also support the development of an indigenous tech sector in the rest of Southeast Asia. Researchers and developers at AI Singapore built SEA-LION (South East Asian Languages In One Network); a family of home-grown Large Language Models (LLMs) that were designed from scratch to better understand the specific languages and dialects of Southeast Asia.
Existing LLMs display a large amount of cultural bias, especially in terms of cultural values, political beliefs, and social attitudes. Many of the most widespread LLMs have been developed in the United States and China – the two giants of AI innovation – and reflect the cultural values of each country. The development of an indigenous Southeast Asian LLM is designed to cater to the region’s many disparate social groups, helping to lower the bar for AI adoption among government, business, and individuals in the wider region. More broadly, by supporting AI adoption, an indigenous LLM such as SEA-LION can enable would-be entrepreneurs to use their new digital skills to develop their own domestic tech sectors. While the scale of adoption potential within Singapore is more limited than other countries in the region, as a first mover Singapore is shaping AI strategy and regulation as a true thought leader.
All of this regional growth is made possible through the nexus of Singapore; it is the country’s existing digitally skilled workforce, well-developed tech ecosystem, and access to startup capital that enabled AI Singapore to embark upon a programme such as SEA-LION.
Source: SEA-LION (2025)
Version 1.
Outperformed most models based on SEA-HELM benchmarks.
Version 2.
Improved performance on SEA tasks while maintaining credible performance on standard English benchmarks.
Version 2.1.
Enchanced conversational abilities across SEA languages with more contextually appropriate responses.
Version 3.
Outperforms similar-sized open-source models and some larger models in both general and SEA capabilities.
Future Plans.
Release of SLM models, release of larger models, covering more local languages/dialects, better regional alignment and guardrails, multimodality.
AI will accelerate R&D in the healthcare sector.
From 1950 to 2010, the number of new drugs approved per billion US dollars has halved roughly every nine years – meaning, the average cost to develop each drug has doubled. While in the 1950s, it might have cost a few million to develop a new treatment, today it can cost close to S$4.02 billion (US$3 billion) to develop each drug. This has acted as a crucial brake on world growth: while semiconductors have seen exponentially declining costs, this has been offset by the slower innovation in medicine.
There are likely to be multiple causes for this so-called ‘Eroom’s Law’ of cost increases, including the increasing complexity of the drug targets being investigated, the challenge in finding new treatments that perform better than those that already exist, increases in wider regulatory costs and the general decline in research productivity seen across the sciences.13
We have now seen exciting evidence that AI could help offset some of these factors, slowing increases in costs or even reducing them.
Singapore’s biomedical sector has seen significant growth since the creation of the National Biomedical Sciences initiative in 2000. The industry is now seen as a potential fourth pillar for the economy after electronics, chemicals and engineering, with four of the world’s top 5 pharmaceutical companies locating manufacturing facilities in the country. Meanwhile, the number of local biotech companies has risen from 7 in 2012 to over 50 by 2022.14 Singapore is becoming a centre for many of the leading trends in the sector, including biologics, mRNA and RNA-based drug development, as well as precision medicine and genomic science.
Using AI to better understand and simulate how different biological proteins and molecules work.
There are more than 1060 potential drug-like molecules and until now, most drugs have been developed by slow and expensive trial and error.15 The discovery phase for new drugs can therefore take between 4 to 5 years – and even after this, around two-thirds of drugs go on to fail in Phase II trials.
However, new AI driven simulation tools such as Google DeepMind’s AlphaFold make it possible to digitally simulate how different molecules interact with each other. This can cut the time to discover new lead candidates, in some cases from years to weeks or even days.
By integrating AI tools into pharmaceutical R&D, we estimate that Singapore’s biotech firms could reduce the average time for drug discovery by over 40%. This can help reduce the overall cost of drug development while also reducing the time it takes to deliver cures to the people who need them.
Using AI to reduce regulatory and administrative costs.
Almost as significant as the costs to find new drug candidates are the ongoing administrative costs as new drugs go through the lengthy process of clinical trials to ensure their safety. AI tools can support here, helping cut time and support with routine data analysis or clinical paperwork.16 In total, we estimate that AI could augment between 10 to 20% of tasks done by those working on medical R&D. This can help reduce the overall costs of medicine development for pharmaceutical companies, ultimately reducing the cost of medicine for consumers.
Using AI to track the real world impact of drugs once they are being used more widely.
Even after this testing, many drugs have unexpected impacts – both positive and negative – once they are used by the population at large. AI can help monitor for and identify these kinds of effects, particularly if combined with wearables and personal devices that give ongoing, real-time health metrics.
AI-Assisted Structural Analysis in Parkinson’s Disease Research.
Researchers at the Agency of Science, Technology, and Research (A*STAR) and the National Neuroscience Institute (NNI) have applied artificial intelligence (AI)-driven tools, including AlphaFold, to study how the immune system may contribute to the progression of Parkinson’s Disease (PD).
The study investigated how autoantibodies in some PD patients target a neuroprotective protein known as STIP1, potentially interfering with its normal function in the brain. A critical challenge in this research was visualising and interpreting the complex 3D structure of the protein and its interactions with autoantibodies—traditionally a time-intensive process.
To support their structural analysis, the team used AlphaFold, a deep learning-based protein structure prediction tool developed by Google DeepMind, to model how autoantibodies might bind to STIP1 at the molecular level. The use of AI-based protein structure prediction complemented laboratory methods and contributed to a more efficient understanding of disease mechanisms.
These findings may support future efforts in early detection and targeted therapeutic strategies for neurodegenerative diseases.
This research was jointly led by scientists from the A*STAR Singapore Immunology Network (A*STAR SIgN) and the National Neuroscience Institute (NNI), including Prof Tan Eng King, Prof Olaf Rötzschke, Dr Chao Yinxia, Dr Jackwee Lim and Dr Jolene Tan Su Yi.
The application of AI tools such as AlphaFold has transformative implications for the health sector. By reducing the time and cost associated with drug discovery and development, these tools have the potential to counteract Eroom’s Law, which describes the declining efficiency of pharmaceutical research and development. The ability to rapidly generate accurate protein structure predictions can accelerate the development of novel diagnostic tools and therapeutic interventions for a range of diseases, including neurodegenerative disorders like Parkinson’s Disease.
Source: Interview with A*STAR and NNI
AI will help to mitigate the effects of an ageing population.
In the 60 years since its founding, the median age in Singapore has more than doubled from 17 to 36 years. Over the next 60, it is expected to increase again to 57 years – while the share of the population aged over 65 will more than triple.
This demographic shift presents significant challenges for Singapore, as the country will need to find the resources to maintain the quality and expand the provision of health and social care for its older residents, while being able to draw from a shrinking pool of working-age adults to fund and staff these services.
According to government estimates, Singapore needs to hire at least 6,000 new nurses every year in order to keep up with the country’s ageing population, and maintain the quality of healthcare services provided to the country’s residents.17
AI is not a silver bullet for ageing, and it cannot offset all the increased pressures of an ageing population. However, it can ease this transition.
Using AI to offset a shrinking workforce.
To start, the higher labour productivity created by AI can aid in offsetting the naturally lower economic growth that would otherwise arise as a result of a smaller working age population. Beyond this, AI can help industries maintain their existing productivity despite experiencing a shrinking workforce by taking over more repetitive tasks.
Using AI to support diagnosis.
AI can help to diagnose conditions in patients by supporting doctors as part of a routine check up or during a formal diagnostic process, but can also get ahead of future health problems by tracking and identifying subtle changes in personal health data. After being combined with data from your smartphone or watch, AI could help provide earlier detection of emerging health conditions when they are still treatable. For example, in Singapore we estimate that AI could help save over 1,100 lives from earlier detection of chronic conditions such as hypertension, diabetes mellitus or hyperlipidaemia.
Hospitals in Singapore are already employing this technology to help in the diagnostic process. The National University Hospital (NUH) uses home-grown AI to help reduce the amount of time it takes a radiologist to interpret an MRI scan, helping to free up radiologists’ time to focus on other priorities.18
Singaporeans are mostly supportive of the principle of involving AI in the diagnostic process – with greater confidence if a human doctor is still involved in the process.
of Singaporeans would support the use of AI to diagnose patients
of Singaporeans would support the use of AI to diagnose patients, as long as this was overseen by a human doctor
Over the next 60 years, the median age in Singapore is expected to increase to 57.
Using AI to boost healthcare capacity.
AI also can help improve the productivity of the health sector more directly. AI tools can help generate efficiency savings for healthcare workers by taking over administrative tasks, and freeing up more time for them to spend directly interacting with patients. In our modelling, we found that over 30% of health sector workers are likely to have their jobs augmented by AI.
Using AI to support eldercare.
With a rapidly ageing population, and an increasing number of older people living alone, loneliness is likely to become an increasingly pressing problem for Singapore’s elderly population.19 Busy workers may struggle to find time to visit their older relatives, leaving many without sufficient social interaction and contributing to worsening mental health conditions. While the government is in the process of boosting mental health provision by training up to 28,000 volunteers to help older people in their community, AI can also offer people another source of social interaction and to keep older people mentally stimulated.20
Singaporean company Dex-Lab has developed a humanoid robot powered by AI to provide social connection to dementia patients, supporting them as their condition worsens and helping to keep them mentally stimulated – vitally important in slowing the onset of conditions such as dementia. The company partnered with Bright Hill Evergreen Home in Punggol for a trial of the technology, with ‘Dexie’ guiding dementia patients in the care home through daily exercises. Alongside the direct social and health benefits of an AI companion, Dexie also helps to free up human time that would otherwise be spent on the task, allowing care workers to focus on more intensive priorities that would be difficult for technology to replicate.