What does "exposed to AI" mean?
The ILO broke 427 occupations of the international classification (ISCO-08) into 3,265 tasks, and scored each from 0 to 1 by how much of it generative AI could take on with current tools.
An occupation's category does not come from the average alone. It combines the average of its tasks and their spread, around a threshold of 0.5. An occupation with a low average but a few tasks clearly above the threshold is classed "exposed, gradient 1". That is why the tool shows the category and the task detail, never a single percentage.
Exposure measures technical potential. It is neither an employer's decision nor a date by which a job disappears.
Why can an exposed job still be hiring?
An exposed task can be sped up, assisted or reorganized. Because most occupations include tasks that require human input, the ILO concludes that transforming jobs is the most likely impact of generative AI.
The useful question becomes: who is hiring for this job right now? The tool answers with live openings from Brevy's catalogue, read directly from company career pages.
And when you apply, a resume that uses the posting's own words gets more replies: 53% more interview callbacks in a ResumeGo study of 7,287 real applications.
Which jobs are most exposed?
According to the ILO, clerical occupations remain the most exposed, as in its 2023 study. Some strongly digitized occupations have become more exposed, as generative AI now handles specialized tasks.
For each major family, the bar shows the share of its occupations classed as exposed. The count is of occupations, not people: a rare occupation weighs as much as a very common one.
| Occupation family | Occupations | Exposed (gradients 1 to 4) | Minimal exposure | Not exposed |
|---|---|---|---|---|
| Clerical support workers | 29 | 28 | 1 | 0 |
| Professionals | 92 | 39 | 26 | 27 |
| Technicians and associate professionals | 82 | 28 | 20 | 34 |
| Managers | 30 | 6 | 20 | 4 |
| Service and sales workers | 39 | 7 | 8 | 24 |
| Elementary occupations | 33 | 2 | 1 | 30 |
| Plant and machine operators, and assemblers | 39 | 1 | 6 | 32 |
| Craft and related trades workers | 65 | 1 | 2 | 62 |
| Skilled agricultural, forestry and fishery workers | 18 | 0 | 0 | 18 |
ILO 2025 data, count by Brevy. Families: ISCO-08 major groups.
Method and sources
Sourced figures, with their limits.
Where do these figures come from?
Scores and categories are the ILO's: Generative AI and Jobs: A Refined Global Index of Occupational Exposure (Gmyrek et al., Working Paper 140, 2025), published under CC BY 4.0. Brevy does not recompute them.
Search relies on the titles of the European ESCO classification, which links nearly 3,000 everyday occupations to the international classification.
The number of openings is computed when you search, over the live openings in Brevy's catalogue whose title matches the one you picked.
Limits
The 2025 scores were largely produced by AI models, then checked against a survey of workers and expert panels. They are estimates, not measurements.
The index is global: it describes an occupation's tasks, not a country or a company. The classification groups different jobs under one code: "web developer" and "multimedia designer" share the same result.
The opening count only covers Brevy's catalogue, and matches openings by the words in their title. None of these figures predicts what will happen to any particular job.