How is AI changing the job market?
Public studies mostly measure exposure, meaning how much of an occupation's tasks AI could do. The ILO (2025) puts one in four workers worldwide in an occupation with some exposure and expects most jobs to be transformed rather than removed. We found no study that lists jobs that will disappear. For a job search, the practical step is to compare the tasks in each job ad with your own experience.
Public studies rank jobs by how much of their work AI could do. That is a measure of exposure, not of lost jobs. Here is what each study counts and what it leaves out.
In this article
This post explains what public studies count when they write about AI and jobs, and what they do not count. It makes no prediction of its own about which jobs will go.
Which jobs does AI change?
The studies we read rank occupations by exposure, not by loss. Exposure means how much of an occupation's tasks AI could carry out. The ILO's 2025 update says one in four workers worldwide is in an occupation with some generative AI exposure. Its refined index says clerical occupations continue to have the highest exposure.
The IMF takes a wider view. In a January 2024 post it says almost 40% of global employment is exposed to AI, and about 60% in advanced economies. Exposure there covers jobs that AI could help as well as jobs where it could take over tasks.
Does exposure mean my job will be replaced?
No. Both organisations say so. The ILO says most occupations include tasks that still need human input, so transformation is the more likely result than removal. It scores exposure from surveys of workers, expert discussions and AI predictions, and the score reflects perceived potential, not measured use.
The IMF splits the exposed jobs in advanced economies in two. It says roughly half may gain productivity from AI. In the other half, AI may do key tasks now done by people, which could lower labour demand, wages and hiring. It adds that some of these jobs may disappear in the most extreme cases, and that the net effect is difficult to foresee.
How much do the numbers differ between countries?
A lot, and the studies measure different things. The ILO refined index puts 34% of employment in high-income countries in an exposed occupation, against 11% in low-income countries. Only 3.3% of employment worldwide sits in its highest-exposure group. In high-income countries that group holds 9.6% of female employment and 3.5% of male employment.
The IMF figure of about 60% for advanced economies and the ILO figure of 34% for high-income countries should not be set side by side. The IMF counts jobs AI could affect in either direction. The ILO scores tasks that generative AI could automate.
Is AI already changing who gets hired?
There is some early evidence from France, and it comes with limits. The Banque de France compared hiring on permanent contracts from January 2019 to March 2026. For 21 to 29 year olds, hires in the most exposed fifth of occupations were about 28% below their November 2022 level in March 2026, against about 8% in the least exposed fifth. The gap narrows for people aged 30 to 49.
The authors find no statistically significant divergence in total employment between firms that use AI intensively and others. They write that the first observed changes show up more in hires and separations than in a visible fall in the number of employees. They also list other causes that hit the same occupations, such as tighter monetary policy and the post-Covid correction in tech. This is one country and one method, so it does not settle the question for the UK or elsewhere.
What does this mean for a job search?
The studies work at the level of tasks, so the useful comparison is the same one. Read the job ad, list the tasks it names and check them against what you have done. The guide to tailoring your CV to a job description describes the steps.
None of the sources we read says AI changes how a CV should look. How software reads your CV is a separate question, and the guide to an ATS-friendly CV covers it. If you are unsure which field to aim at, our post on how to know what job to look for is a better starting point than any exposure ranking.
How should I read the next headline about AI and jobs?
Check four things before you trust a number.
- What is measured: exposure, tasks, hiring or job losses.
- Who is counted: all workers, one country or one age group.
- The date. The studies above run from 2024 to March 2026.
- The caveats. Good studies state them, as the ILO, the IMF and the Banque de France do.
Applying for a role that is changing?
CVTailor reads the job ad and shows a fit score and the keywords your CV is missing. It writes a new summary for each ad from your experience, and you can edit everything before download.
Frequently asked questions
Is there a list of jobs AI will replace?
We found no public study that gives one. The ILO and the IMF rank occupations by exposure, and both say exposure does not tell you how many jobs will be lost.
Why do the ILO and the IMF give different shares of exposed jobs?
They count different things. The IMF (2024) counts jobs that AI could affect, whether helped or harmed. The ILO (2025) scores occupations by how many of their tasks generative AI could automate. The numbers are not meant to be compared.
Are entry-level jobs affected more than others?
One study we read points that way for France. The Banque de France found that hiring of 21 to 29 year olds on permanent contracts fell more in the most exposed occupations. Its authors say other causes could also explain it, and it covers one country.
Should I mention AI skills on my CV?
We found no source on this. If the job ad asks for experience with specific tools, the tasks you have done with them belong on your CV in your own words.
Sources
- The ILO says one in four workers worldwide are in an occupation with some degree of generative AI exposure, scores about 30,000 tasks, and says most jobs are likely to be transformed rather than eliminated because human input is still needed. ILO, Generative AI and jobs, 2025 update
- The ILO refined index puts 34% of employment in high-income countries and 11% in low-income countries in an exposed occupation. It reports 3.3% of employment worldwide in the highest-exposure group, and 9.6% of female against 3.5% of male employment in that group in high-income countries. It says exposure is not job loss. ILO, Generative AI and jobs, refined global index (2025)
- An IMF blog post (January 2024) says almost 40% of global employment is exposed to AI, about 60% in advanced economies, 40% in emerging markets and 26% in low-income countries. In advanced economies, roughly half of the exposed jobs may benefit and the other half may see lower labour demand. It calls the net effect difficult to foresee. IMF Blog, AI will transform the global economy (2024)
- The Banque de France compared hiring by age and occupation, January 2019 to March 2026. It reports that hires of 21 to 29 year olds on permanent contracts fell more in the most AI-exposed occupations, and that total employment in firms using AI intensively shows no statistically significant divergence. The authors list other possible causes. Banque de France, Generative AI and young people (2026)