The BaseResearchInternational Labour Organization
Generative AI and Jobs: A Refined Global Index of Occupational Exposure
An ILO working paper, written with Poland's research institute NASK, updates the ILO's 2023 index of how far generative AI could perform each occupation's tasks. Scoring tasks by survey, expert review and AI models, the authors find 24% of global employment has some exposure, 3.3% the highest, and judge transformation of jobs likelier than replacement.
Open the source ↗Top story of the week
Why it matters
The same thing, for different readers.
What it means for you
About one in four workers is in an occupation where AI could do some tasks, clerical work most; the authors think changed jobs likelier than lost ones.
For lawmakers
The figures are an upper bound on possible effects, not job-loss forecasts; the authors urge social dialogue and targeted policy, and find women and richer countries more exposed.
For researchers
Scores blend a survey of 1,640 Polish workers, expert review and AI models; all tasks are weighted equally, and task-level scores are public for reuse.
The question it raises
If a job is a bundle of tasks, automating some can reshape work rather than end it, though the tasks left over may make a job unsustainable or undesirable.
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“Globally, one in four workers are in an occupation with some GenAI exposure. 3.3% of global employment falls into the highest exposure category, albeit with significant differences between female (4.7%) and male employment (2.4%).”
Found word for word in the source
“Indeed, the results indicate that few jobs consist of tasks that are fully automatable with current GenAI technology; nearly all occupations have some tasks that require human input.”
Found word for word in the source
“Clerical occupations continue to have the highest exposure levels. Additionally, some strongly digitized occupations have increased exposure, highlighting the expanding abilities of GenAI regarding specialized tasks in professional and technical roles.”
Found word for word in the source
“It is important to stress that these new estimates reflect “exposure” to GenAI and not the actual impact on employment. As in 2023, the 2025 estimates represent the upper threshold of the percentage of employment that could be affected if GenAI technology were fully implemented.”
Found word for word in the source
“Finally, the estimates do not account for new jobs that may be created, nor the technological advances that could potentially automate more tasks in the future. In that sense, our estimates present a static snapshot of the exposure of existing occupations to GenAI at the start of 2025.”
Found word for word in the source
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- A task-level exposure index. A web survey of employed people in Poland on an online research platform (1,535 retained of 1,640; 52,558 task ratings) scored a sample of 2,861 of the 29,753 tasks in the Polish occupational classification; experts reviewed 608 tasks; two AI models (GPT-4o and Gemini Flash 1.5) reconciled survey and expert scores; an AI assistant then predicted scores for all tasks in the international ISCO-08 classification. Occupations are sorted into four exposure gradients by the mean and spread of their task scores, and the ILO global employment model turns that into shares of employment by sex, region and income group.
- What it found
- 24% of global employment is in an occupation with some generative AI exposure (gradients 1 to 4): 3.3% in the highest gradient (4.7% of female and 2.4% of male employment) and 20.5% in the lower three. Exposure is 34% of employment in high-income countries and 11% in low-income ones; in high-income countries 9.6% of female and 3.5% of male employment is in the highest gradient. Clerical jobs stay the most exposed, with data entry clerks, typists and bookkeeping clerks among 13 occupations in the top gradient. The 2023 estimate was 2.3% in its automation category and 29.2% across augmentation and an unknown middle category.
- What it does not show
- Exposure is a score for existing tasks, not employment impact: the authors call the estimates an upper threshold, say they leave out new jobs and later advances, and describe them as a static snapshot at the start of 2025. Scores rest on a Polish survey (76% of respondents never or rarely used generative AI), expert judgment and AI-model arbitration, weight every task equally, and treat Poland as an upper-threshold setting. Exposure does not say whether workers will be replaced or retrained; the authors think transformation likelier but warn that partial automation could still cut demand for some jobs.
- Status
- A working paper: not yet peer reviewed
The record
- Questions
- What happens to work, income and the economy?
- People
- Paweł Gmyrek (Wrote it, Senior Researcher, Research Department, ILO) · Janine Berg (Wrote it, Senior Economist, Research Department, ILO) · Karol Kamiński (Wrote it, Senior Researcher, Analysis and Research Department, NASK-PIB) · Filip Konopczyński (Wrote it, Director, Analysis and Research Department, NASK-PIB) · Agnieszka Ładna (Wrote it, Manager, Analysis and Research Department, NASK-PIB) · Balint Nafradi (Wrote it) · Konrad Rosłaniec (Wrote it, Researcher, Analysis and Research Department, NASK-PIB) · Marek Troszyński (Wrote it, Expert, Analysis and Research Department, NASK-PIB, and Assistant Professor, Civitas University, Warsaw)
- Also covered by
- International Labour Organization
- What we read
- We read the whole source.
- Found by
- A research pass: work, income and the economy, gathered October 13, 2026, via
research-work(tier 1: the source’s own channel or an official record) - Cite this
- International Labour Organization, “Generative AI and Jobs: A Refined Global Index of Occupational Exposure”, May 20, 2025. Entry in the Ctrl AI Base, https://ctrlai.com/base/ilo-2025-05-20-genai-jobs-refined-exposure-index
Our notesText: the ILO's HTML edition of Working Paper 140 (Read online), read in full apart from the annex tables. The paper itself gives only the year (copyright 2025); the day, 20 May 2025, comes from the ILO publication page, which also lists Balint Nafradi among the authors; the paper's byline and author notes name seven people, and Nafradi is included here on the publication page's word, without a title. Quotations use the first author's id. Scores partly come from AI models (the authors say the writing and calculations are their own). The World Bank report in this pass builds its jobs shares on this ILO-NASK index, and the IMF note in this pass cites the ILO's 2023 study; the editor may want to relate them.