Will AI take jobs?
Generative AI is already changing tasks in many office jobs, especially clerical work, but so far there is no clear sign of falling overall employment. The main open questions are how fast the change will come and whether people early in their careers lose out.
Exposure is not the same as job loss
Most research on AI and work measures exposure: how many of the tasks in an occupation a generative AI system could, in principle, help with or carry out. A high exposure score means the job is likely to change. It does not mean the job will disappear.
The ILO's refined global index from May 2025 (Working Paper 140) finds that about one in four workers worldwide is in an occupation with some exposure to generative AI, but only 3.3% are in the highest exposure group. The ILO concludes that transformation of jobs is the most likely outcome, because most occupations still require human input. The IMF estimated in 2024 that almost 40% of jobs worldwide, and around 60% in advanced economies, are exposed to AI. In advanced economies it expected roughly half of the exposed jobs to benefit through higher productivity, while the other half could see lower demand for labour, fewer hires and pressure on wages.
Who is most exposed
Clerical and administrative work stands out in every major study. In the ILO index, clerical occupations have the highest exposure. Because women hold more of these jobs, 4.7% of employed women globally are in the highest exposure group, compared with 2.4% of men. Exposure is also far higher in high-income countries (34% of employment) than in low-income countries (11%).
For Sweden, the AI Take My Job analysis, which combines ILO exposure scores with employment data from Statistics Sweden (SCB) for 2024, finds that 13% of employees work in occupations with significant or the highest exposure, and 29% in exposed occupations overall. Among administration and customer service clerks, 56% are in the two highest groups. Women are slightly more affected than men (13.5% versus 12.6% in the two highest groups).
What has been observed so far
The International AI Safety Report 2026, written by more than 100 experts, finds no visible effect of AI on overall employment so far. New research from the US and Denmark found no link between an occupation's AI exposure and its total employment. Several other studies, however, point to falling employment for early-career workers in the most exposed occupations since late 2022, while older workers in the same occupations have held steady or grown.
Usage data shows how AI is used in practice. Anthropic's Economic Index (January 2026) found that 52% of conversations on its consumer chatbot were augmentation, where a person works together with the AI, and 45% were automation, where the AI does the task. In business use through the API, about three quarters was automation. Use is still concentrated: roughly a third of consumer conversations concern programming.
What employers expect
In the World Economic Forum's Future of Jobs survey (2025), employers expect new jobs equal to 14% of today's employment to be created by 2030, and jobs equal to 8% to be displaced, a net increase of 7%. These are expectations rather than measurements, and they cover all major trends, not only AI. The largest absolute declines are expected among clerical and secretarial workers, such as cashiers and administrative assistants. 40% of employers expect to reduce staff where AI can automate tasks, while 85% plan to prioritise upskilling.
Economists disagree about the longer term. Some expect modest effects on total employment. Others argue that wages and employment could fall significantly if AI becomes able to do nearly all tasks more cheaply than people.
What is being done
The ILO and IMF both point to the same tools: social safety nets, retraining and policies that make it easier to move between jobs. The IMF stresses that policy choices will shape whether AI widens or narrows inequality. In the EU, the AI Act classifies AI used in recruitment and other employment decisions as high-risk, with requirements on data quality, documentation and human oversight (see the page on bias and discrimination). A practical difficulty, noted by the International AI Safety Report, is that productivity gains and risks to workers often come from the same applications.
What you can do
- Look up your occupation on this site to see which tasks are most exposed, and which parts of your work rely on judgement, relationships or physical presence.
- Learn to use AI tools in your current role; 85% of employers in the WEF survey plan to prioritise upskilling.
- If you work in administration or customer service, build skills that complement AI, such as quality control, complex customer contact and knowledge of how your organisation works.
- Take part when your workplace decides how AI is introduced; agreements on training and transition make a difference.
How exposed is your job?
Search 423 occupations and see which tasks AI could do.
Sources
- Generative AI and Jobs: A Refined Global Index of Occupational Exposure (ILO Working Paper 140) – International Labour Organization (ILO), 2025
- AI Will Transform the Global Economy. Let's Make Sure It Benefits Humanity – International Monetary Fund (IMF), 2024
- The Future of Jobs Report 2025 – World Economic Forum, 2025
- Anthropic Economic Index report: Economic primitives – Anthropic, 2026
- International AI Safety Report 2026 – International AI Safety Report (chair: Yoshua Bengio), 2026
Data updated: 2026-11-19