AI Is Already Reshaping Jobs in Asia — IMF Data Shows Demand Down 10% in Exposed Roles
The IMF’s October 2026 outlook finds employer demand down nearly 10% in AI-exposed occupations across 19 Asian economies. But the story is not simple displacement — and Indonesia is named as a country that must act.
OzeroPublished October 18, 20265 min read
The International Monetary Fund’s October 2026 Regional Economic Outlook for Asia and the Pacific delivers one of the clearest measurements yet of AI’s impact on work: drawing on roughly 40 million online job postings across 19 Asian economies, employer demand has fallen by nearly 10 percent in occupations more exposed to AI, relative to less-exposed ones, since ChatGPT launched in November 2022.
The headline number — and the nuance
A 10% demand drop sounds like a displacement story, but the IMF stresses it is not that simple. Although AI capabilities have advanced considerably since 2022, the decline in demand for workers in highly exposed occupations appears to have stabilised. The initial shock has levelled off — which suggests labour markets are adapting, not collapsing.
One reason for the stabilisation: entirely new employment categories are emerging. Employers increasingly post roles requiring prompt engineering, retrieval-augmented generation (RAG), and AI agent development — skills that barely existed as job titles three years ago. Workers who can integrate AI into business processes and translate its capabilities into practical outcomes are in growing demand.
Adoption is uneven — and that matters
- Only around 6% of firms in Asia explicitly request AI skills in job postings — adoption is still early and patchy.
- Firms that HAVE adopted AI show stronger labour-demand growth than non-adopters, including in occupations highly exposed to automation.
- That last point is key: inside AI-adopting firms, the technology appears to complement human workers even as it automates specific tasks.
Why Indonesia is named
The report singles out India, Indonesia, and the Philippines — countries with growing working-age populations that need to create huge numbers of productive jobs, just as AI transforms the very sectors expected to absorb those workers. For Indonesia, the IMF’s prescription is education reform, reskilling programmes, and digital infrastructure so that AI complements job creation rather than blocking it.
This is the real policy challenge: it is not about stopping AI, but about how fast workers, businesses, and education systems adapt to it. Countries that reskill quickly turn the 10% figure into a transition statistic; countries that do not risk turning it into a structural one.
“The future of workspaces in Asia will depend not simply on how rapidly AI advances, but on how effectively economies, businesses and workers adapt to it. — IMF, October 2026”
Ozero’s take
For anyone building in crypto or tech in Indonesia, this report is a signal, not a threat. The stabilisation of the demand decline plus the emergence of AI-integration roles means the premium is shifting from “doing the task” to “orchestrating the AI that does the task.” Prompt engineering and AI agent skills are not hype terms anymore — they are literally showing up in millions of job postings. Learn them now while adoption is still at 6%.
Risks to watch
- Labour-market data describes averages across 19 economies — individual sectors and roles vary widely. Treat the 10% figure as a directional signal, not a prediction for any specific job.
Conclusion
AI has already reshaped Asian labour demand — that part is measured, not speculated. What happens next depends on adaptation speed. The workers and countries that learn to work with AI, rather than compete against it, are the ones the data favours.