Why hyper-efficient Singapore is Asia’s most AI-anxious society

(Edited and refined by Candice Chan, Grace Chong and Josephine Hong, with the assistance of AI translation.)

A survey of four Asian markets has found that Singaporean businesses are the most proactive in promoting artificial intelligence (AI) training and providing AI-related resources, while Singaporeans are also the most likely to pay for AI tools. At the same time, among the four markets, respondents in Singapore expressed the greatest concern about being replaced by AI and becoming overly reliant on it.

Academics interviewed said the findings reflect the fact that AI has demonstrated tangible value in Singaporean workplaces, prompting more people to incorporate it into their daily work. However, amid the rapid advance of AI capabilities, people are also feeling more acutely the pressure brought about by the technology.

Between May and June this year, the 2026 AI Survival Anxiety Survey was conducted by Lianhe Zaobao, Taiwan’s Global Views Research and Malaysia’s Sin Chew Daily. The survey covered respondents aged 18 and above in Singapore, Malaysia, Taiwan and mainland China, collecting more than 2,700 valid responses, including 444 from Singapore.

The survey found that AI has become an increasingly common tool in the daily lives of Singapore respondents. Nearly four in five (79.5%) use AI at least once a week, while 57.4% use it every day. Only 4.3% said they had never used AI, underscoring the technology’s widespread adoption in Singapore.

Among Singapore respondents, 80.4% use AI to improve work efficiency. Other common uses include learning and skills development (61.5%) and seeking health-related information (31.3%). Respondents aged between 18 and 34 recorded the highest rate of using AI to improve work efficiency, at 98%.

By contrast, AI was used less frequently for entertainment, creative pursuits, casual conversation or emotional support, suggesting that Singaporeans are more inclined to use the technology for practical purposes.

One in four pays for AI tools

Singapore also recorded the highest proportion of respondents paying for AI tools out of their own pocket, with around one in four (24.8%) subscribing personally. Another 18.9% said the costs were covered by their employer or educational institution, while 54.3% relied solely on free versions.

The corresponding proportions of respondents paying for AI themselves were 24.3% in mainland China, 16.8% in Malaysia and 14.9% in Taiwan.

Among Singapore respondents, those aged 18 to 34 were the most likely to pay personally for AI services, with a subscription rate of 37.3%. By educational attainment, 37.6% of respondents holding a master’s degree or above paid for AI tools, significantly higher than university graduates (21.6%) and those with a secondary-school education or below (3.6%).

The survey also showed that Singaporean companies are the most active in promoting AI adoption. Some 46% of respondents said their workplaces provided AI training programmes or AI-related tools and resources, compared with 30.5% in mainland China, 29% in Taiwan and 28.5% in Malaysia.

At the same time, Singapore respondents estimated that AI currently helps them complete an average of 28.9% of their work tasks, second only to mainland China (31.5%). Among them, 18.9% said AI was already capable of completing more than half of their workload.

Highest anxiety over AI replacement

As AI becomes more deeply embedded in the workplace, employees are also becoming increasingly concerned about their future roles.

The survey found that Singapore respondents recorded the highest levels of anxiety among the four markets over being replaced by AI and struggling to remain competitive in the workplace, scoring 5.82 and 5.19 respectively on a 10-point scale.

Singapore respondents also expressed concern about becoming increasingly dependent on AI. Their greatest worry was that being unable to access AI would significantly affect their work or daily lives. They registered an anxiety score of 5.10, the highest among the four markets.

Although AI is already widely used at work, Singapore respondents were relatively conservative about its potential to boost their earnings. Only 22% believed their income would increase because of AI over the next three years, lower than the corresponding figures in Taiwan (36.5%) and Malaysia (29.8%). Another 42.4% said it was still too early to judge the impact of AI on their income.

Employers must help workers build skills

Professor Ben Leong, director of the AI Centre for Educational Technologies at the National University of Singapore, told Lianhe Zaobao (LHZB) that Singapore’s relatively high rate of paid AI subscriptions shows that AI is clearly useful for the sort of work that is common in Singapore, so that many users are willing to pay for more advanced features.

Asked why greater AI adoption appears to coincide with higher workplace anxiety, Leong said one reason is that today’s AI seems to produce very high quality work without needing much input, making users more aware of the pace of technological progress and the pressure it brings.

However, he stressed that AI is still far from replacing human workers entirely.

He said AI cannot do everything perfectly, and human input is still needed.

He drew the analogy of “copilot” and “autopilot” mode in using AI. In the copilot model, the human remains critical while AI makes work better and faster. The real risk of replacement arises under the autopilot model, where the human outsources almost all of the work to AI.

To ease anxiety, Leong argued that, besides improving public understanding of AI’s capabilities and limitations, companies and public-sector organisations should do more to help employees build up their skills.

Varying preferences for AI tools

Preferences for AI tools differ markedly across the four Asian markets. In Singapore, Malaysia and Taiwan, users primarily rely on US-developed and other international AI tools. In mainland China, however, domestic platforms dominate.

However, when assessing the overall strength of AI ecosystems, respondents in Singapore and Malaysia rated mainland China more highly for its AI infrastructure and adoption of AI applications. Across all four markets, however, respondents agreed that the US leads in semiconductor technology, computing power, AI model development, and the policy and regulatory environment.

According to the survey, around 70% of respondents in Singapore and Malaysia primarily use US-developed and international AI tools such as ChatGPT, Gemini and Claude. In Taiwan, the figure is even higher, at 88.6%.

Mainland China presents a stark contrast. More than half of respondents (52.8%) primarily use AI tools developed by Chinese tech companies, including Qwen, ERNIE Bot and Doubao, while only 11.7% mainly rely on US-developed and international tools.

AI capabilities ranked

Beyond AI tool usage, the survey also included the US and asked respondents to compare the AI capabilities of different countries and regions in AI infrastructure, AI model development, AI adoption, semiconductor technology and computing power, as well as policy and regulation.

Respondents in Singapore and Malaysia generally viewed mainland China as the leader in AI infrastructure and the deployment and adoption of AI applications. Respondents in mainland China and Taiwan, by contrast, ranked the US first.

Across all four markets, respondents were unanimous in placing the US at the top for AI model development, semiconductor technology and intelligent computing power, reflecting broad recognition of its leadership in core technologies and high-end hardware.

The same consensus extended to AI policy and regulation, with respondents in all four markets ranking the US first. Notably, respondents in mainland China, Singapore and Taiwan all placed Singapore second, highlighting recognition of the country’s strengths in AI governance and regulatory frameworks.

The findings also underscore the distinct strengths of China and the US in the AI landscape. The US continues to lead in AI model development and advanced computing power, with tools developed by companies such as OpenAI, Google and Anthropic dominating international markets.

China, meanwhile, has in recent years stepped up efforts to build a robust domestic AI ecosystem. Driven by tech companies including Alibaba, ByteDance and Baidu, it has rapidly expanded the deployment of AI applications and their integration across industries.

Respondents in Singapore gave consistently balanced ratings across the different dimensions of AI development. Of the six indicators, semiconductor technology and intelligent computing power received the highest score (5.92), followed by AI adoption (5.89) and the policy and regulatory environment (5.87).

To advance the country’s AI development, the Singapore government announced in February the establishment of the National AI Council, chaired by Prime Minister Lawrence Wong, to oversee the national AI strategy. It will also support businesses in adopting AI through the National AI Impact Programme, while building a workforce equipped with both domain expertise and AI capabilities.

As AI-generated deepfake videos, images and audio become increasingly widespread, being misled by such content tops the list of ten AI-related fears for Singaporeans.

Fear of scams tops all

According to the survey, Singapore respondents recorded an anxiety index of 7.01 out of ten regarding the fear of “deepfakes and frauds”, ranking highest among the survey’s ten AI-related fears. Across the four Asian regions surveyed, Singapore’s level of anxiety was second only to Taiwan’s, surpassing both mainland China and Malaysia.

The survey also found that people across all four regions share concerns over the growing difficulty of distinguishing authentic content from AI-generated material. Singapore respondents expressed anxiety over “assessing reliable content” and “identifying AI-generated information”, with scores of 5.90 and 5.61 respectively. On both measures, Singapore again ranked second only to Taiwan.

By contrast, Singapore respondents showed the least anxiety about “not knowing how to learn AI”, which scored 3.81 points — the lowest among the ten concerns.

Deepfake technology is already being weaponised by fraudsters. In May this year, the Singapore Police Force uncovered a deepfake video scam in which fraudsters used AI to forge a Zoom meeting featuring Singapore President Tharman Shanmugaratnam, Prime Minister Lawrence Wong and several ministers, defrauding a single victim of at least S$4.9 million (US$3.8 million).

Highest trust in official sources

Leong noted that these concerns most likely stem from the frequent occurrence of scam cases in recent years. “Most people personally know of people who got scammed. I know of two people who were scammed out of hundreds of thousands of dollars,” he said.

He emphasised that this does not imply that Singaporeans are more gullible. Rather, it reflects Singapore’s highly digitalised society and fast-paced environment, where people often process information rapidly and can fall prey to scams during a momentary lapse in vigilance.

The survey further indicated that as deepfake content becomes more pervasive, the public leans more heavily on official and professional sources for reliable information. An overwhelming 93.9% of Singapore respondents placed their highest trust in information released by the government or official agencies, followed by reputable media outlets (87.2%), and professionals or research institutions (73.0%).

In stark contrast, trending social media content (4.1%), online influencers (1.6%), and messages forwarded by friends (1.6%) ranked as the least trusted sources of information.

This article was first published in Lianhe Zaobao as “亚洲四地调查：新加坡职场最积极推动AI应用 用户焦虑感也最高”, “四地AI工具偏好分化 新马台多用美系 中国大陆依赖本土平台” and “AI深伪内容越来越逼真 本地民众最怕被误导”.

Related: US-China rivalry undermines fight against cyber scams in Southeast Asia | AI Stefanie, scams and fake news: China acts on AI regulation