The International Labour Organisation has released a comprehensive study revealing that generative artificial intelligence will have far-reaching implications for the ASEAN workforce, potentially affecting nearly 80 million workers across the region's 11 member states. The findings, drawn from ILO analysis of employment patterns and AI exposure across occupational categories, paint a nuanced picture of technological disruption—one characterised by substantial exposure to AI technologies but, crucially, an absence of widespread job losses at this stage.
According to the ILO's 2025 projections, approximately 22.9 per cent of total employment in ASEAN—translating to roughly 80 million workers—is concentrated in occupational roles that face at least some degree of potential exposure to generative AI. This figure is striking in its scale, suggesting that nearly a quarter of the regional workforce operates in sectors where AI capabilities could meaningfully alter the nature of daily tasks, skill requirements, or employment structures. However, the study introduces an important distinction that tempers initial alarm: the proportion of workers facing the most severe exposure is considerably smaller. Only 3.3 per cent of the ASEAN workforce, approximately 11.7 million employees, occupies roles classified as having the highest exposure to generative AI disruption.
The research underscores that the current trajectory does not support catastrophic employment scenarios. Despite the significant exposure figures, there is presently no empirical evidence of large-scale redundancies triggered by generative AI adoption. This finding carries particular weight for policymakers across Southeast Asia who have expressed concern about technological unemployment. The report notes that employment in highly exposed occupations has continued to expand, suggesting that labour market dynamics remain resilient even as AI capabilities develop. The critical caveat, however, is that generative AI adoption itself remains in early stages across most of ASEAN, with implementation concentrated primarily in technology-intensive sectors rather than dispersed throughout the economy.
Geographical variation within ASEAN reveals significant disparities in both AI exposure and readiness. Singapore emerges as the regional leader, with 42.2 per cent of its workforce employed in occupations showing more than minimal AI exposure. This reflects the city-state's position as a global financial and technology hub, where digital infrastructure and tech-oriented employment are deeply embedded. The Philippines follows at 28.1 per cent, a figure that the ILO attributes partly to the nation's substantial service and information technology sectors. Indonesia, Southeast Asia's largest economy, records 21.7 per cent exposure, while Vietnam registers 20.8 per cent and Thailand 20.6 per cent. These figures illustrate how economic structure and sectoral composition directly influence vulnerability to AI-driven labour market change.
A particularly notable finding concerns the gendered dimension of AI exposure. Women are more than twice as likely as men to work in occupations classified as having high generative AI exposure. This disparity reflects long-standing occupational segregation patterns, with women concentrated in clerical, administrative, and professional roles—precisely those positions where AI tools show considerable capability for task augmentation or replacement. The concentration of women in these roles introduces a policy imperative for ASEAN governments: any effective response to AI disruption must address the asymmetric risk burden falling on female workers, lest technological change exacerbate existing gender-based economic vulnerabilities.
Interestingly, age does not appear to be a significant differentiator in AI exposure. Young workers aged fifteen to twenty-four and adult workers demonstrate broadly comparable levels of potential exposure to generative AI capabilities. This finding contradicts some narratives suggesting that younger cohorts face systematically higher vulnerability. Instead, it suggests that the nature of occupational roles—rather than worker demographics per se—drives exposure levels. An administrative assistant aged fifty-five faces similar AI-related pressures as a twenty-five-year-old in an identical role, though their capacity to retrain or adapt may differ substantially.
The report identifies what it terms a significant "preparedness gap" across ASEAN, with capacity and readiness varying markedly between member states. Singapore stands apart as a globally competitive AI ecosystem, combining sophisticated digital infrastructure, abundant technical talent, and a coordinated government strategy for AI integration. Other ASEAN nations, by contrast, face considerable obstacles in building comparable institutional and human capital capacity. This divergence threatens to create a two-tier regional landscape where technology adoption reinforces existing economic disparities, with advanced economies like Singapore capturing productivity gains while less-developed neighbours struggle to manage transition challenges.
GenAI adoption patterns reveal another critical insight: uptake remains concentrated in technology-intensive occupations despite administrative and office roles displaying exceptionally high potential exposure. This disconnect between exposure and actual implementation suggests a lag between technological capability and organisational deployment. Many businesses, particularly in traditional sectors or smaller enterprises across the region, have not yet integrated generative AI into their operations, even where such integration would theoretically be beneficial. This implementation gap creates a window of opportunity for proactive policy intervention before disruption becomes acute.
The ILO's analysis demonstrates that approximately 67 per cent of ASEAN employment continues in occupations with no identified exposure to generative AI. These roles—concentrated in agriculture, manual labour, service sectors requiring physical presence, and specialised technical fields—represent a form of relative insulation from AI disruption. However, this figure should not breed complacency. Agricultural employment, while currently protected from direct AI displacement, faces indirect pressures through AI-driven optimisation of supply chains and resource allocation. Moreover, as AI capabilities expand beyond current generative tools, exposure categories may broaden significantly.
The ILO's recommendations for ASEAN pivot around four interconnected priorities. First, human-centred governance frameworks must guide AI integration, ensuring technology serves worker wellbeing rather than simply maximising corporate efficiency. Second, inclusive skills development programmes—expanded upskilling and reskilling initiatives with particular attention to women and youth—become essential public investments. Third, micro, small and medium enterprises require targeted support to overcome financial and technical barriers to AI adoption, preventing a situation where only large corporations benefit from productivity gains. Fourth, strengthening regional knowledge exchange and coordinating human resource development strategies enables ASEAN member states to learn collectively and avoid duplicative investment in policy experimentation.
For Malaysia specifically, these findings carry profound implications. As a middle-income economy with a diversified employment base spanning manufacturing, services, technology, and agriculture, Malaysia likely experiences exposure patterns intermediate to regional leaders like Singapore and developing economies like Indonesia. The country's substantial skilled workforce in finance, petrochemicals, electronics, and business services positions many Malaysians within moderate-to-high exposure categories. Policymakers must act proactively to ensure that AI-driven transformation enhances rather than diminishes employment quality and opportunity.
The broader regional context suggests that ASEAN faces not an imminent labour market catastrophe but rather a critical period of transition management. The window remains open for deliberate policy choices that could shape whether generative AI becomes a tool for inclusive prosperity or concentrated advantage. Member states that move decisively on skills development, social protection enhancement, and inclusive innovation ecosystems may navigate the transition successfully. Those that delay risk widening inequality gaps both within nations and across the region.
