Singapore is calling on businesses to be measured in their approach to artificial intelligence (AI) and look closely at the impact of AI to determine how they should leverage the technology.
The Singapore government has been keen to drive AI adoption amongst its population and businesses, introducing initiatives to guide key verticals, including healthcare, in their AI transformation and offering support to help workers adapt.
However, moving ahead doesn’t mean moving blindly.
“We are at a technological inflexion point, learning to work and coexist with AI,” said Ong Ye Kung, Singapore’s Minister for Health (MOH) and Coordinating Minister for Social Policies, during his speech at NCS Impact 2026.
“We cannot charge ahead driven solely by commercial considerations, even as recursive AI systems gain self-reinforcing intelligence, agency, and influence,” Ong said.
“Otherwise, the machines just seem wiser than their makers,” he said. “We must be wiser, more humanistic, and practical. We must decide deliberately where to embrace AI, where to rein it in, and where human judgement and effort must prevail.”
He noted that while AI can now read diagnostic scans, health regulations still require clinicians to review the AI-powered analysis and provide a diagnosis.
AI, too, can produce movies, but most people still prefer films shot with real scenery and human actors, he added.
AI’s impact depends on several factors, such as the type of industry, how demand and market forces evolve, behaviour of companies, and the nature of the job.
For instance, Ong believes consumers will continue to value craft and authentic human creation, limiting the extent of AI application in these areas. In Japan, there is major pushback against using AI to create anime cartoons, he noted.
There will, however, be jobs or tasks that can be substituted by AI, he said.
Routine and process-intensive work, such as data collation and report preparation is under threat, he explained, adding that drivers are concerned about the prospects of AI-driven autonomous vehicles.
Such worries of those affected should be addressed, the minister said, pointing to the government, unions, and employers to support workers in the transition.
Through agencies such as the Skills and Workforce Development Agency, the Singapore government will provide training programmes and mitigate job uncertainties, as well as “persuade” employers to “keep an open mind”, Ong said.
“For certain jobs under threat, governments may need to decide whether regulations and guardrails are necessary,” he said.
He highlighted that autonomous vehicles in China are required by local laws to have a safety officer sit behind the wheel, ready to intervene in an emergency. Aircrafts, too, already can be flown largely on autopilot, but are mandated by international aviation regulations to have a full pilot crew.
Human oversight remains essential in safety-critical systems, while work still should be primarily carried out by humans in sectors where human trust and empathy are paramount, he said.
Find AI tools that actually work
As it is, Singapore professionals apparently are amongst the least sceptical about AI, but use the technology least globally.
Some 29% in the city-nation describe themselves as AI sceptics. This figure is lower than the global average of 37%, according to a study released by Salesforce, which polled more than 1,500 desk workers across 15 markets, including India, Germany, Japan, Italy, and Australia.
In comparison, in countries such as the UK, the US, and France, 53% of respondents describe themselves as AI sceptics.
Despite the lower number of sceptics, though, just 6% of Singapore desk workers say AI is a core part of their daily work, ranking them amongst the lowest across the board. The global average clocks at 11%.
There may be reasons for the low adoption rate.
In Singapore, 31% of workers have experienced unsuccessful AI pilots, of which 40% point to generic outputs as a reason for failure, the study revealed.
Another 38% highlight low trust in outputs as a reason for unsuccessful pilots, while 30% cite results that lacked business context.
The findings suggest Singapore workers are not hesitant to adopt AI, but are instead held back by tools that fail to meet expectations of relevance, accuracy, and reliability, Salesforce said.
“Singapore workers are not standing in the way of AI — they’re waiting for AI that works for them,” said Paul Carvouni, Salesforce’s senior vice president and general manager for Asean. “While workers’ enthusiasm towards AI is a headstart, poor pilots are leaving real business potential on the table.”
“[Organisations] have to move past generic tools and use AI that is trusted, grounded in business context, and built into daily work,” Carvouni added.
Amongst more than 500 respondents globally who are able to move from pilot to daily use, most point to success factors such as role-specific training, integration of AI into existing workflows, and strong data security.
Build the right infrastructure
According to Ong, good use of AI should encompass a strong digital operating environment and good quality data, with sound policy and organisational structure guiding its use.
He said Singapore’s healthcare sector already has been working to strengthen all three components, where MOH is in the final phases of replacing legacy IT systems with integrated systems across the sector.
This includes the adoption of a sector-wide common electronic medical record (EMR) system by 2028, he added.
A cloud-based platform also has been built to securely pull together healthcare data, analytics tools, and AI capabilities, he said.
Called Healix, the platform enables public healthcare institutions to develop, train, test, and deploy AI solutions, he added.
It also will better facilitate the use of LLMs (large language models) to power AI tools, he said.
And with healthcare already a well-regulated industry, it has the time and space to consider how best to move ahead with AI, Ong said.
“AI in healthcare should never be like the proverbial hammer looking for nails. Instead, we take a use case approach, identifying problems and areas of improvement, and using technology to address them,” he said.
MOH will scale AI use cases that are effective and impactful, he noted, adding that the ministry currently is working on 10 such system-wide initiatives, which include automating medical note-taking and clinical coding and predicting outpatient attendance.
Another initiative, called Singapore Medical Foundational AI Model (SIMFONI), aims to build AI models that support clinicians, such as suggesting possible diagnoses and treatments.
While such models already exist, they are trained using patient data and medical guidelines in other countries, Ong said.
He noted that SIMFONI’s models will be developed and trained using Singapore’s clinical practice guidelines and local clinical data, contextualised to the local population.
The foundational AI model will start by focusing on cardiometabolic diseases and eye diseases, and will be deployed system-wide when ready, he said.
AI diagnostics, though, cannot fully replace radiologists.
This is necessary to maintain clinical skills as well as ensure human judgement remains relevant when it comes to caring for another human being, he noted.
Avoid the pitfalls
AI should be a multiplier, not a replacement, of human expertise, said NCS CEO Sam Liew at the conference.
Amongst other priorities, AI-led enterprises need to reskill their workforce with AI capabilities and equip them with the necessary tools to enhance their work, Liew said.
NCS itself is “rewiring” how it operates to become an AI-led service organisation, he said.
Its efforts have included establishing a team that focuses on the company’s AI initiatives, such as building reusable platforms, establishing safe adoption practices, and driving the necessary skillsets.
The systems integrator now runs on a new operating structure that focuses on 10 industry-specific groups, with two service organisations as the underlying foundation, including applications and communications engineering for AI agent-led delivery.
Liew further noted that all NCS employees are paired with at least three AI agents to support their tasks.
The vendor also is offering a playbook that is designed to guide organisations in their AI deployments, he said, adding that the document draws on insights from more than 100 AI projects.
It includes advice on how companies should go about planning their AI roadmap and pitfalls to avoid. It lists common causes of failed AI projects, such as unclear cost structures, unchanged work processes, and data that lacks context.
The AI playbook also urges companies not to underestimate the cost of token consumption and to reward outcomes, rather than use.
