The Impact of Generative AI on Southeast Asia’s Workforce
Fears of a new industrial revolution driven by artificial intelligence replacing manual labor have not yet materialized in Southeast Asia. However, white-collar workers are being urged to prepare for potential changes as the region grapples with the effects of generative AI (GenAI). According to a July report by the International Labour Organization (ILO), nearly one in four workers in the region face GenAI disrupting or affecting their jobs.
Clerical, administrative, and professional positions are most at risk, according to the report. Manual trades, craft, and agricultural work remain relatively insulated, with the effects of AI varying across the region based on economic development, digital infrastructure, and labor policies.
Being most “exposed” to GenAI does not mean that entire professions will be replaced—just that some could move up the value chain or evolve, analysts say. Phu Huynh, senior employment specialist at the ILO, noted that understanding automation and its impact on jobs has become more nuanced than in early forecasts surrounding the Fourth Industrial Revolution.
GenAI exposure among craft and trades workers and skilled agricultural workers is generally low, with only 3.3 per cent of Southeast Asian workers in occupations with the highest exposure category. Around 67 per cent had jobs with no identified exposure to GenAI. These patterns suggest that while GenAI has the potential to affect a sizeable share of workers across ASEAN, occupations with the highest levels of exposure still account for a relatively small share of total employment.
Nearly 80 million women and men across the region are employed in jobs exposed to GenAI. There were twice as many women (4.8 per cent) as men (2.3 per cent)—reflecting the concentration of women in clerical, administrative, and selected professional occupations. The gender gap was most pronounced in Thailand and the Philippines, where women are around three to four times more likely than men to work in highly exposed occupations.
Singapore has the highest share of workers exposed to GenAI—at 42.2 per cent—which the brief said might partly reflect the country’s highly knowledge-intensive occupational structure. The Philippines follows at 28.1 per cent, reflecting the country’s service-oriented economy and prominence of its information technology and business process management sector. Indonesia (21.7 per cent), Vietnam (20.8 per cent), and Thailand (20.6 per cent) also recorded significant levels of exposure.
Transforming Jobs
Despite the scale of AI exposure, there was limited evidence of large-scale job losses associated with GenAI. In ASEAN, widespread labor market disruption is not yet visible, but continued monitoring of job impacts is critical, especially among young workers in countries showing slower growth or declining employment in selected entry-level administrative and clerical jobs.
Across the region, GenAI remains dominant over agentic AI and is used for tasks such as drafting, translation, coding, and clerical and administrative work. GenAI operates on prompts to create content or data, while agentic AI can plan and execute deeper steps towards a goal with higher autonomy.
The emerging consensus is that the impact of this technological transition will be greater on the transformation of jobs and tasks, rather than outright displacement.
“GenAI is likely to reshape white-collar jobs because many routine cognitive tasks can be automated or augmented. However, the key issue is less about jobs disappearing and more about jobs evolving,” said Huynh. “Ensuring that workers have the skills to use AI effectively will be critical to supporting economic upgrading and sustaining the region’s transition up the value chain.”
ILO’s findings mirror similar research conducted on India and Indonesia by Aapti Institute. According to Nighat Nighat, an associate at the India-based institute, AI systems are likely to amplify demand for advanced digital and cognitive skills while reducing the demand for routine and clerical activities. “Also in greater demand will be skills that complement GenAI, such as critical thinking, creativity, and emotional intelligence.”
Entry-level white-collar roles were “precisely the tasks GenAI is best at automating,” she said, with workers and experts pointing out a decline in demand for simpler, repetitive tasks.
Readiness Factor
Readiness across the board remains a primary concern. Young workers in advanced economies potentially face higher exposure because a larger share of jobs involves tasks that can be augmented or automated by AI. In developing economies, lower adoption may reduce immediate disruption, but it can also limit access to the related productivity gains from innovation and new opportunities.
According to the report, Singapore was best positioned to harness AI opportunities due to its strong digital infrastructure and ecosystem, and whole-government approach. Malaysia, Thailand, Brunei, Indonesia, the Philippines, and Vietnam were next. At the bottom tier are Cambodia, Laos, Myanmar, and East Timor, which “generally exhibit lower levels of readiness, reflecting more limited digital infrastructure, institutional capacity, and technological capabilities.”
Analysts also stressed the importance of organizations supporting their workers through the shifts. Ben Teehankee, president of the Responsible AI Council of the Analytics and Artificial Intelligence Association of the Philippines, noted that the adoption of newer GenAI and related technologies is dependent on the attitudes and belief systems of managers towards the effectiveness and cost-efficiency of these technologies.
AI adoption could widen the gap between high-skill, well-paying jobs and low-skill, low-paying jobs, Nighat said. “When GenAI skills are mature enough to match or exceed human ability, they tend to substitute workers; but when jobs involve more complex workflows, disruption tends to be minimal.” This creates a two-tier effect: workers doing routine tasks get substituted, facing downward wage pressure, while those whose jobs are AI-complemented gain productivity and see their skills become more valuable and in greater demand.
Policy reforms and country support could help the transition process, Teehankee said. “Without significant policy reforms and government intervention, broad-based AI automation will tend to impact workers in advanced economies first but this will cascade to the value chains in developing countries, as companies look for cost efficiencies to stay competitive.”
The key determinant, Huynh said, would be “how effectively countries support workers and enterprises to prepare for and navigate the transition, including investments in skills, digital infrastructure, and inclusive and ethical AI adoption.”
