Hiring demand in the occupations most exposed to artificial intelligence has fallen nearly 10 percent relative to less-exposed roles across 19 Asian economies since ChatGPT launched in November 2022, the International Monetary Fund said in its October 2026 Regional Economic Outlook for Asia and Pacific. The finding draws on roughly 40 million online job postings. The IMF said the decline has stabilised even as AI systems grew more capable, and framed the shift as work being reshaped rather than eliminated outright.
The gap opened as AI crossed capability thresholds, the report said. Demand in more-exposed occupations had dropped by nearly 10 percent by the point systems could handle tasks such as finding and fixing a bug in a small software library, work that would take a person about an hour. The decline then held steady even as models learned to complete jobs that would take a person several hours.
Alongside the fall, the IMF recorded a rise in demand for workers who can use the technology. Postings now seek skills such as prompt engineering, retrieval-augmented generation and the building of AI agents, which have spread since 2023. About 6 percent of firms in Asia explicitly ask for AI skills, a sign that adoption remains uneven. Companies that adopt AI are showing stronger hiring growth than those that do not, including in highly exposed occupations, which the IMF reads as evidence the technology can complement workers even as it displaces specific tasks.
The fund was careful about what the data can and cannot show. Job postings measure new recruitment demand, not actual hiring or total employment, and employers can shift how and where they advertise. The IMF said reduced hiring was the stronger force in the evidence it examined, but that fuller effects on employment, wages and hours would emerge only with further analysis.
The implications diverge sharply by country. India, Indonesia and the Philippines must create enough productive jobs for growing working-age populations while AI transforms the sectors expected to absorb them. The IMF noted that emerging Asian economies may gain short-term insulation because many workers remain in agriculture and other less-exposed sectors, but warned this could become a structural trap if AI makes the climb toward higher-productivity manufacturing and services harder. Its recommended response runs beyond adoption to education reform, reskilling, digital infrastructure and stronger labour-market institutions.
India's own debate illustrates the difficulty. A Times of India analysis argued that the country's AI-skilling push, including the 60,000-crore-rupee PM-SETU scheme to upgrade 1,000 government training institutes, needs a demand test: the Periodic Labour Force Survey found only 4.2 percent of people aged 15 to 59 had received formal vocational training. The report's finding that AI-exposed hiring has stabilised rather than collapsed is the more reassuring half of the picture. Whether that holds as models take on longer and more complex tasks is the open question the IMF cannot yet answer.