Download
Since OpenAI launched ChatGPT, the AI chatbot, in November 2022, employers across Asia have cut demand for workers in jobs more exposed to AI. The IMF’s October 2026 Regional Economic Outlook: Asia and Pacific offers some of the first empirical evidence of this shift, finding that employer demand in AI-exposed occupations fell by nearly 10% relative to less-exposed jobs.
The International Monetary Fund’s report is based on 40 million online job postings from 19 low-income, emerging and advanced Asian economies. This lets it track what skills employers want, by economy, industry and occupation, month by month through April 2026.
The IMF compared jobs where AI is more useful with jobs where it is less useful. Coding is an example of a highly exposed job. Managing supplier relationships is the example of a less-exposed one. The comparison was made within the same economy, industry, and month, and it accounts for the scope for remote work. This helps separate AI’s effect from other forces, such as post-pandemic hiring shifts or earlier over-hiring.
So, are jobs disappearing?
Yes, but not exactly; hiring has slowed, as per the IMF. Employer demand has fallen by nearly 10 per cent in more-exposed occupations relative to less-exposed ones. Before ChatGPT’s launch, hiring in both groups moved together. Demand changed little until mid-2023, then weakened measurably in 2024 and 2025.
One limitation in going by current trends, as per the report, is that job postings only capture new hiring and related roles. The IMF says the full weight of the shortfall will be felt gradually, as it builds up over time.
Is it getting worse as AI gets smarter?
The IMF links hiring trends to how long a task AI can complete would take a person. When ChatGPT launched, AI could handle things like simple arithmetic, which takes a person seconds. There was little impact on jobs.
By the time AI could find and fix bugs in a small code library, a task of about an hour for a person, demand in exposed jobs had dropped nearly 10 per cent.
Story continues below this ad
AI can now take on day-long tasks, such as training a machine learning model. Yet the jobs picture has not worsened. The report says employers still see a need for workers, even where AI is most useful.
The answer lies in a new kind of demand. The IMF tracked “emerging AI skills” such as prompt engineering, retrieval-augmented generation and building AI agents. These meant little before 2023 but have spread quickly since. They show up more often in the more-exposed occupations, and demand for them is rising fastest there.
In simple terms, employers want people who can use AI well, build with it, and connect it to their company’s own knowledge and processes.
40 mn
job postings studied
~10%
demand fall in AI-exposed jobs
~6%
of firms ask for AI skills
Hiring demand since ChatGPT
Less-exposed jobs (e.g. supplier ties)Baseline
More-exposed jobs (e.g. coding)Nearly 10% lower
Does smarter AI mean fewer jobs? Tap to see
Rising demand for new AI skills
Prompt engineering
Retrieval-augmented generation
Building AI agents
Firms asking for AI skillsAbout 6%
Pulling down
Less hiring where AI is useful. Stronger so far.
Pushing up
New openings for AI-skilled workers. Growing.
Little effect so far in lower-income economies.
Express InfoGenIE
Who is driving this?
A small group of firms. About 6 per cent of firms in Asia ask for AI skills in their job postings. Demand for workers is growing faster at these AI-adopting firms than at firms that do not ask for such skills, even in more-exposed occupations.
Story continues below this ad
That also tells us something else. Most firms have not yet adopted AI, which points to real barriers in fitting the technology into existing workflows and processes.
Two forces, pulling in opposite directions
The IMF says AI’s overall effect reflects two forces at once: less hiring where AI is more useful, and new openings for workers who can put it to use.
For now, reduced hiring is the stronger of the two, which is why demand has fallen in exposed jobs. But the offsetting opportunities have already begun to appear, and they appear to be growing over time.
The impact also differs across the region. The report finds little effect so far in lower-income economies, so AI’s footprint varies widely from place to place.
Story continues below this ad
So far, the data does not show AI wiping out jobs. It shows a shift in what employers ask for. In exposed occupations, hiring is thinner, but the people who can work with AI are in growing demand. That puts the focus on skills and on how quickly workers and firms can adapt.