Skip to main content
AI in Asia
Pan-Asia

SEA Workers Fear AI Over-Dependence Most

Milieu Insight's six-market study finds 53% of SEA workers list AI over-dependence as their top concern, ahead of privacy and job loss.

· Updated Apr 26, 2026 7 min read
SEA Workers Fear AI Over-Dependence Most

Southeast Asian Workers Are More Worried About AI Over-Dependence Than Job Loss, And That Tells Us Where The Region Is Headed

A new Milieu Insight study of 3,000 workers across six Southeast Asian markets has produced a result that should reframe how regional employers, regulators, and platform operators think about AI deployment in 2026. Over-dependence on AI ranks as the top concern across all six surveyed markets at 53%, ahead of privacy concerns at 40% and job loss at 34%. That is a fundamentally different concern hierarchy from the one driving most Western AI policy debate, and it has direct implications for product design, training programmes, and regulatory framing across the region.

What The Survey Actually Says

Milieu Insight ran the survey across Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam. The full findings are summarised in the TechKnowNews report and were also covered in regional press in late April 2026. The headline result is that workers in every single one of the six markets ranked over-dependence as their top AI concern, with only modest variation in the absolute scores.

Vietnamese workers showed the highest overall AI optimism at 66%, followed by Thai workers at 58%. Singapore workers, despite having the most mature AI infrastructure and policy environment in the region, sat slightly below the regional optimism average. The full pattern points to a region that is eager to use AI but deeply self-aware about the risk of losing skills, judgement, or independence in the process.

Asian workers are not afraid of AI replacing them. They are afraid of AI making them less capable. That is a meaningful distinction with real product and policy implications.

Gerald Ang, Chief Executive, Milieu Insight

Why The Concern Hierarchy Matters

Western AI policy debate has been dominated by job displacement narratives, privacy framing, and bias concerns. Those concerns matter in Asia too, but the Milieu data shows they are not the top issue for working-age Southeast Asians. The dominant concern is more existential and more cultural. It is the worry that easy AI access will erode the discipline of thinking through hard problems, the patience to learn skills slowly, and the social trust that comes from professional competence built over years.

That concern shows up in product design preferences. The same survey found that 67% of respondents preferred AI tools that "explain their reasoning" over tools that "give the fastest answer." A similar majority preferred tools that "ask clarifying questions" over tools that "produce immediate output." Those preferences track exactly with the Socratic-mode AI tutor designs that are now standard in Singaporean and Hong Kong universities, and they explain why AI assistance products that emphasise scaffolding tend to retain Southeast Asian users better than products that emphasise speed.

By The Numbers

53%

53% of Southeast Asian workers list AI over-dependence

53% of Southeast Asian workers list AI over-dependence as their top concern, ahead of privacy and job loss

40%

40% list privacy as their top concern

40% list privacy as their top concern, the second-ranked issue

34%

34% list job loss as their top concern

34% list job loss as their top concern, the third-ranked issue

66%

66% of Vietnamese workers report optimism about AI

66% of Vietnamese workers report optimism about AI, the highest rate in the six-market survey

58%

58% of Thai workers report optimism

58% of Thai workers report optimism, the second-highest rate

3,000

3

3,000 respondents across Indonesia, Malaysia, the Philippines, Singapore, Thailand, and Vietnam

How Employers Are Responding

The most interesting employer responses have been at the platform companies that operate across multiple Southeast Asian markets. Grab, Sea Limited, and Gojek all run significant internal AI training programmes that include explicit modules on what the companies call "AI capability stewardship," which is a polite way of saying training employees to keep their underlying skills sharp even as AI tools handle routine work.

Grab's 13 AI experiences launch, which made the company the most visible regional adopter of agentic AI in March 2026, was paired with a structured upskilling programme designed to keep customer-facing staff competent at handling exceptions that the AI systems escalate. Sea Limited's AI Centre of Excellence has a parallel mandate. Both companies have learned that AI deployment in Southeast Asia works better when paired with explicit anti-deskilling programmes.

Foreign multinationals operating in the region are gradually adapting. Microsoft's 150,000 Thai workers AI certification programme explicitly emphasises sustained skills development rather than tool familiarity. Google's expanded Singapore research footprint includes a new emphasis on training and capability programmes targeted at early-career engineers in the region. NEC's Japan partnership with Anthropic for 30,000 employees has a similar stewardship dimension built into the deployment plan.

The companies that will retain Southeast Asian workforce trust through the agentic AI shift are the ones that pair tool deployment with deliberate capability stewardship. The data is clear.

Lily Kong, President, Singapore Management University

What This Means For Regulators

Southeast Asian regulators have generally been more measured than their Western counterparts on AI restriction. The Milieu data suggests why. With job displacement not the dominant concern, the political pressure for restrictive AI legislation is weaker than the global headline narrative implies. The harder regulatory question is consumer protection, especially around products that might encourage over-reliance for vulnerable users including students, older workers, and those in the informal economy.

That framing aligns with the Singapore ISO/IEC 42119-8 testing standard work, which emphasises evaluation methodology rather than substantive use restriction. It also aligns with the Vietnamese AI legislation now under development, which leans toward consumer-protection framing rather than competition-policy framing. Expect a wave of similar consumer-focused regulatory work across Indonesia, the Philippines, and Malaysia through 2026.

ConcernIndonesiaMalaysiaPhilippinesSingaporeThailandVietnam
AI over-dependence54%52%53%51%54%54%
Privacy41%39%42%40%38%40%
Job loss33%35%32%36%33%35%
AI optimism52%54%51%53%58%66%

For broader pan-regional context, see our coverage of the Stanford AI Index findings on Asia AI optimism and the Asia cross-border AI talent flow analysis.

The AIinASIA View: The Milieu data is the most important regional sentiment finding of 2026 to date. Southeast Asian workers are not anti-AI. They are pro-AI but worried about losing themselves in the process, which is a more sophisticated and policy-actionable concern than the simple displacement framing dominant in Western coverage. Companies and governments that build their AI deployment strategies around capability stewardship will outperform those that focus narrowly on productivity gains. Expect that to become the default framing across ASEAN by Q3 2026, with Singapore, Vietnam, and Thailand leading the policy and product response.

Frequently Asked Questions

Why is over-dependence the top concern in Southeast Asia rather than job loss?

The pattern likely reflects a combination of cultural emphasis on skill development, a more flexible labour market that has absorbed past technology shifts, and a population that has lived through the rapid rise of mobile and platform technologies and learned to be wary of dependency.

Does the same pattern hold in North Asia?

Partial data from Japan and Korea suggests similar over-dependence concerns but with stronger weighting toward privacy and elder-care framing. The Milieu study did not cover those markets, but parallel surveys from local sources point in the same broad direction.

What do these results mean for AI tool design?

Tools that emphasise scaffolding, reasoning visibility, and clarifying questions tend to perform better with Southeast Asian users than tools that emphasise raw speed. Product teams targeting the region should weight retention metrics over time-to-first-output.

Will this shift Southeast Asian AI regulation?

It already has, indirectly. Singapore's standards work, Vietnam's draft legislation, and the broader ASEAN governance discussion all reflect consumer-protection and stewardship framing rather than displacement-driven restriction.

How should multinationals adapt their training programmes?

Pair tool rollouts with explicit capability stewardship programmes that train employees to handle the exceptions, judgement calls, and edge cases that AI systems escalate. The companies doing this well, including Grab, Sea, and Microsoft Thailand, are already running visible internal programmes along these lines.