Job ads for AI-exposed work fell nearly 10 percent, study finds

Employers are posting fewer job ads for work that AI can do well. A new study from the International Monetary Fund (IMF) found that hiring demand in highly AI-exposed jobs fell by nearly 10 percent compared with less exposed jobs. At the same time, ads asking for new AI skills are growing.

What the study found

The researchers looked at more than 700 million online job postings across 110 economies. They found that demand in highly exposed occupations weakened relative to other roles during 2023 and 2024. The drop approached 10 percent toward the end of that period. Averaged over the whole period, the drop was smaller, at around 6 percent.

One detail may calm some nerves. The decline largely leveled off once AI could complete tasks that take a person about an hour. It did not keep getting worse as the technology improved. The study does not say the risk to jobs has passed.

What it does not show

The study counts job ads, not layoffs. It does not show how many people lost work or how many are employed overall. Its long-term effects on jobs also remain uncertain, since they depend on how businesses adopt AI and how workers respond.

Who it affects

Advanced economies saw stronger hiring declines in highly exposed jobs. The effect was smaller in emerging market economies and statistically insignificant in low-income economies.

Job type mattered too. Where AI can work alongside people, the decline was about 4.7 percent. Where there is little room for that, it was about 9.5 percent.

New skills in demand

The researchers found 51 new AI-related skills showing up in job ads. Demand for them rose quickly from 2023, especially in jobs already exposed to AI.

• Prompt engineering
• AI agents
• ChatGPT and Microsoft Copilot
• Retrieval-augmented generation and vector databases

Among more than 560,000 firms, those asking for these skills generally saw faster growth in job postings than those that did not.

What to do

For workers, the study suggests skills that pair well with AI may matter more. For employers, the task is to work out which jobs AI can automate, where it can help current staff, and what new skills the business will need.