Singapore is asking businesses to evaluate their approach to artificial intelligence (AI) and closely monitor the impact of AI to decide how to leverage the technology.
The Singapore government is keen to accelerate the adoption of AI in the country. population and companiesI will introduce Initiatives to guide key areasincluding healthcare, is working on AI transformation and helping workers adapt.
However, if we proceed This does not mean moving blindly..
“We are at a technological inflection point, learning to work with and coexist with AI,” Singapore’s Minister of Health (MOH) and Coordinating Minister for Social Policy Ong Ye Kung said in a speech at NCS Impact 2026.
“we cannot be charged first Even as recursive AI systems gain self-reinforcing intelligence, agency and influence, they are driven solely by commercial considerations,” Ong said.
“Otherwise, the machine just looks smarter than its manufacturer,” he says. “We have to be smarter, more humane and more practical. Carefully decide where to deploy AIWhere to suppress it, and where human judgment and effort must prevail. ”
He noted that although AI can now read diagnostic scans, health regulations still require clinicians to review AI-powered analysis and provide a diagnosis.
He added that AI can also make movies, but many people still prefer movies shot with real scenery and human actors.
The impact of AI will depend on several factors, including: Industryhow demand and market forces evolve, corporate behavior, and the nature of work.
For example, Ong believes that consumers will continue to value crafts and authentic human creations, which will limit the scope of AI applications in these areas. Major in Japan Backlash against the use of AI He pointed out that it was to create an animated manga.
However, he says there will be jobs and tasks that can be replaced by AI.
Routine, process-intensive tasks like data collation and report generation are under threat, he explained, adding that drivers are concerned about the future of AI-powered self-driving cars.
He said these concerns of those affected should be addressed, noting that governments, trade unions and employers should support them. workers in transition.
The Singapore government, through agencies such as the Skilled Workforce Development Board, will provide training programs to reduce job uncertainty and “persuade” employers to “keep an open mind”, Mr Ong said.
“Governments may need to make decisions about specific jobs that are under threat. Regulations and guardrails “It’s necessary,” he said.
He emphasized that local laws require self-driving cars in China to have safety personnel sitting behind the wheel, available to intervene in emergencies. Although most aircraft can already fly on autopilot, international aviation regulations require a full pilot to be present.
He said that while human oversight remains essential in safety-critical systems, work should still be primarily performed by humans in areas where human trust and empathy are paramount.
Find AI tools that actually work
In fact, while Singaporean experts are clearly the least skeptical of AI, they also have the least use of the technology globally.
About 29% of people living in urban areas say they are AI skeptics. According to one study, this number is lower than the global average of 37%. Survey results released by Salesforcesurveyed more than 1,500 desk workers in 15 markets including India, Germany, Japan, Italy, and Australia.
In contrast, 53% of respondents in countries such as the UK, US and France describe themselves as AI skeptics.
However, despite the small number of skeptics, only 6% of desk workers in Singapore say AI is core to their daily work, ranking it among the lowest overall. The world average clock is 11%.
There may be a reason for the low adoption rate.
In Singapore, a study found that 31% of employees have experienced AI pilot failure, with 40% citing generic output as the reason for failure.
A further 38% highlighted unreliable outputs as a reason for pilot failure, and 30% cited results lacking business context.
This finding suggests that Singaporean workers are not hesitant to adopt AI, but rather because of the tools that prevent them from implementing it. meet expectations Salesforce says it’s more relevant, accurate, and reliable.
“Singapore’s workers are not standing in the way of AI; they are waiting for AI to work for them,” said Paul Carvouni, senior vice president and general manager of Asean at Salesforce. “Employee enthusiasm for AI is a head start, but low pilot competency is undermining real business potential.”
“[Organisations] We need to move away from generic tools and use AI that is trusted, contextual to the business, and embedded in daily operations,” Calvouni added.
Of the more than 500 respondents around the world who were able to move from pilot to daily use, most cited success factors such as role-specific training, integrating AI into existing workflows, and strong data security.
Build the right infrastructure
According to Ong, successful use of AI must include a strong digital operating environment and high-quality data. sound policy and the organizational structure that guides its use.
He said Singapore’s healthcare sector is already working to strengthen all three elements, and the Ministry of Health is in the final stages of replacing legacy IT systems with integrated systems across the sector.
This includes implementing a common electronic medical record (EMR) system across the industry by 2028, he added.
He said a cloud-based platform is also being built to securely integrate medical data, analytical tools and AI capabilities.
The platform, called Healix, will enable public healthcare organizations to develop, train, test and deploy AI solutions, he added.
Also, Using the LLM (Large language models) will power AI tools, he said.
Ong also said that because the healthcare industry is already well regulated, there is time and space to consider how best to move AI forward.
“AI in healthcare Never be like the proverbial hammer looking for the nail. Instead, we take a use case approach, identify problems and areas for improvement, and use technology to address them,” he said.
He said the Ministry of Health plans to expand effective and impactful AI use cases, adding that the ministry is currently working on 10 system-wide initiatives, including automating medical record creation and clinical coding, and predicting outpatient attendance.
Another initiative, called the Singapore Medical Foundational AI Model (SIMFONI), aims to build AI models to support clinicians, including suggesting possible diagnoses and treatments.
Such models already exist, but they are trained using patient data and medical guidelines from other countries, Ong said.
He said SIMMONI’s model will be developed and trained using Singapore clinical practice guidelines and local clinical guidelines. Contextualized data To the locals.
He said basic AI models will start with a focus on cardiometabolic and eye diseases, and will be rolled out across the system as soon as they are ready.
However, AI diagnostics cannot completely replace radiologists.
this is, maintain clinical skills In addition to ensuring human judgment It remains important when it comes to caring for other humans, he noted.
avoid the pitfalls
NCS CEO Sam Liu said at the conference that AI should multiply human expertise, not replace it.
Liu said AI-driven companies must, among other things, prioritize retraining their employees on AI capabilities and providing them with the tools they need to enhance their work.
NCS itself is “rewiring” the way it operates to become an AI-driven service organization, he said.
That effort includes creating a team focused on the company’s AI efforts, including building a reusable platform, establishing safe deployment practices, and driving the necessary skill sets.
The systems integrator now operates with a new operating structure focused on 10 industry-specific groups, with a foundation of two service organizations including applications and communications engineering for AI agent-driven delivery.
Liew further said that every NCS employee is paired with at least three AI agents to support their operations.
Vendors also provide a playbook He said the document is designed to guide organizations in their AI adoption, adding that the document leverages insights from more than 100 AI projects.
It includes advice on how companies should approach planning their AI roadmaps and the pitfalls to avoid. Here we list common reasons why AI projects fail, including opaque cost structures, unchanged work processes, and data lacking context.
The AI Handbook also recommends that companies: Don’t underestimate costs It emphasizes token consumption and rewards results rather than usage.
