Which Countries Have the Most Workers in AI-Exposed Jobs?

How we measured it

WorldJobsData scores each of the nine ISCO-08 occupation groups for AI exposure on a 0 to 10 scale. Two groups score 6 or higher: Professionals (6.5) and Clerical support workers (8.5). Managers and Technicians score 5.5, and every other group scores 3.5 or lower. For each country we divide the number of workers in those two groups by total employment in the country's occupation table.

Employment comes from ILO ILOSTAT (CC BY 4.0) and national statistics offices, including Eurostat and the US Bureau of Labor Statistics. Countries whose data arrives in finer sub-groups, such as Germany and the Netherlands, are rolled up to the nine major groups first, so every country is measured at the same level. Of 206 countries, 173 have all nine major groups in their occupation table. The other 33 are left out because their source tables report only seven or eight groups, so the groups cannot be compared like for like. That list includes China, Japan, Canada, South Africa and Saudi Arabia.

Data years run from 2003 to 2026: 141 of the 173 countries are on 2020 or later data, and 32 are on 2019 or earlier. The year is shown in every table below. The 6/10 cutoff is our choice, and the scores behind it are model-estimated, not measured. See the methodology and the sensitivity check below.

The 15 countries with the highest share

Source: ILO ILOSTAT and national statistics offices, from each country's occupation table in site/data/countries/[code].json; AI exposure scores from site/data/isco_scores.json. Share = Professionals + Clerical support workers as a percentage of total employment in the table. Computed and checked 2026-10-13.

Ten of the top 15 are European. The others are the United States, Israel, Singapore, South Korea and Australia. Luxembourg's figure (51.0%) rests on 2022 data for 302,400 workers, so treat it as the least certain number in the table. The United States has 29.6% of its 144 million workers in professional jobs and 11.3% in clerical jobs, which puts it third of 173. Germany has a slightly higher clerical share (12.4%) but fewer professionals (23.7%), and lands 11th.

The ten largest workforces

Rankings of small countries can mislead about where most workers actually live. The ten largest workforces in the dataset range from 5.5% to 40.9%. China and Japan are not in this table because their data does not meet the nine-group requirement described above.

India has the largest workforce in the dataset (476.6 million workers) and 8.2% of it is in the two exposed groups, which is about 39 million people. The United States has about 59 million. India's share is roughly a fifth of the US share, yet its exposed group is about 66% as large in absolute terms, because its workforce is more than three times bigger.

Income and the exposed share

Sorting 167 countries by GDP per capita (World Bank, US dollars) into four equal groups shows a steep gradient. The rank correlation between GDP per capita and the exposed share is 0.88.

At the bottom of the table, agriculture and elementary work are the largest groups (see below), so the gradient largely reflects how big the farming share is. The data show where the exposed jobs are, not who gains or loses from AI. The gap between the groups is large, but each group has wide spread: the richest quarter runs from 14.5% to 51.0%.

The ten lowest

In all ten, agriculture is the largest occupation group, and agriculture plus elementary work together make up 55% to 89% of employment. In Niger, agriculture is 73.0% of employment and elementary occupations another 3.6%. In Laos the figures are 64.4% and 8.0%, in Tanzania 51.6% and 28.2%. Agriculture scores 3.0 and elementary work 2.0 out of 10 on our scale. Note the age of some of this data: Haiti's table is from 2003 and Burundi's from 2020.

Does the cutoff matter?

The 6/10 line is a judgement, so we checked two alternatives. If managers and technicians (both 5.5) are counted as well, the exposed share rises everywhere, and the leaders become Singapore (73.7%), Luxembourg (72.2%), Switzerland (65.1%), Netherlands (64.7%), United Kingdom (64.4%). The ranking changes little: the rank correlation with the main list is 0.97. If only clerical workers (8.5) are counted, the rank correlation falls to 0.82, and the top three are small territories: Réunion (22.7%), Puerto Rico (16.9%), Aruba (14.4%). The link with national income survives both alternatives: the rank correlation with GDP per capita is 0.89 with managers and technicians added and 0.81 for clerical workers alone, against 0.88 for the main measure.

The bigger caveat is the scores themselves. They were produced by an AI model on 28 May 2026, guided by published research. They are informed judgements, not measurements, so a different model or date could shift them.

What this means

If you work in a country near the top of the table, more of the workforce is in occupation groups where current AI can assist with the most tasks. Near the bottom, the same AI touches a much smaller share of paid work, and the data shows why: those economies employ most people in farming and elementary jobs. Neither situation is a forecast of job losses.

For a single country, the occupation-level breakdown, with wages and the full list of groups, is on the country pages and in the explorer.

Spotted an error or know a better source? Report it and we will check it against the source data.