What is an AI exposure score?

An AI exposure score (0–10) measures how much of a job's day-to-day tasks can plausibly be done by current AI systems. A score of 8–10 means most tasks are highly automatable today. A score of 2–3 means very few tasks are automatable with current technology. Scores are research-based estimates informed by Frey-Osborne (Oxford), OECD and IMF studies, not official government statistics.

ISCO-08 is the International Standard Classification of Occupations (2008 version). It groups all jobs into 9 major categories consistently across every country, allowing fair cross-country comparison. WorldJobsData maps all national occupation data (BLS, ONS, Eurostat, ABS, StatCan) to ISCO-08 so you can compare the US workforce directly against India, Germany or any other country.

Where does the employment data come from?

Employment data comes from official government sources only: ILO ILOSTAT (206 countries), BLS Occupational Outlook Handbook (US, 341 occupations), ONS ASHE (UK), Statistics Canada, ABS (Australia), and Eurostat SES 2022 (EU countries). Wage data comes from the same sources. No data is invented or estimated - all sources are cited on each country page.

How do I compare two countries?

Click the Compare button in the top toolbar, then select a second country from the dropdown that appears. Both countries will be shown side by side using the same colour scale, making it easy to spot differences in workforce composition and AI risk.

Which occupation has the highest AI exposure score?

Clerical support workers have the highest AI exposure score at 8.5 out of 10. This group includes secretaries, administrative assistants, customer service representatives, data entry clerks and bookkeeping clerks. Their core tasks - processing, routing and communicating structured information - are highly automatable by current AI systems.

What is the difference between AI risk and robotics risk?

AI risk measures cognitive automation - how much of a job's information processing, writing, analysis and communication can be done by AI software. Robotics risk measures physical automation - how likely it is that robots and machines will replace manual tasks. A factory operator scores moderate on robotics risk but low on AI risk. A data entry clerk scores high on AI risk but low on robotics risk.

What does WFH potential mean?

WFH potential (0–10) measures how much of an occupation can be performed remotely from home. Occupations requiring physical presence - such as construction, healthcare and food service - score 1–2. Knowledge-based occupations such as software development, finance and legal work score 7–9. Higher WFH potential generally correlates with higher AI exposure.

How accurate are the AI exposure scores?

AI exposure scores are research-based estimates, not official government statistics. They are informed by three major studies: Frey and Osborne (Oxford, 2017), the OECD Future of Work report, and the IMF 2024 Generative AI paper. They represent a reasonable consensus on task-level automation potential. Individual circumstances vary - a score of 7/10 for a profession does not mean 70% of those workers will be displaced.

Help & Glossary | WorldJobsData - AI Job Risk & Workforce Data

Ask a question - type anything about WorldJobsData

What does the color mean? What is Tier A? How do I compare countries? What is an AI score? Why fewer occupations? What is ISCO?

Answer

Occupation Classification

ISCO-08

International Standard Classification of Occupations (2008 version). A global system that groups all jobs into standardised categories so data from different countries can be compared fairly. WorldJobsData uses it as the common language for every country's workforce.

Example: "Managers" is ISCO group 1. "Professionals" is group 2. These groups exist in every country's data.

ISCO Major Group

The broadest level of job classification - 9 groups numbered 1–9 (plus group 0 for Armed Forces). Every country in WorldJobsData shows at least this level. Each rectangle in the treemap is one major group.

Groups: 1 Managers · 2 Professionals · 3 Technicians · 4 Clerical · 5 Service & Sales · 6 Agriculture · 7 Craft & Trades · 8 Machine Operators · 9 Elementary

Employment Share (%)

What percentage of a country's total working population is in each occupation group. The size of each rectangle in the treemap reflects this - bigger rectangle = more workers in that group.

BLS Occupation (US only)

For the United States, WorldJobsData has data for 341 individual job titles from the US Bureau of Labor Statistics Occupational Outlook Handbook. These are much more specific than ISCO groups - e.g. "Nurse Practitioners", "Software Developers", "Welders". Only the US has this level of detail.

Data Tiers

Tier A - Full data

21 countries with occupation-level wage data from official national statistics. Employment, AI scores, gender breakdown, and actual salary figures are all shown. These countries have the richest insight cards.

Countries: US, UK, Germany, Australia, Canada, Netherlands, Spain, Poland, and 13 more EU/EEA countries.

Tier B - Employment only

19 countries where we have occupation employment data from Eurostat but wages were not available or were suppressed. You'll see the treemap, AI scores, and gender data - but no pay figures.

Countries: France, Italy, Sweden, Denmark, Finland, and others.

Tier C - ILO data

166 countries with workforce data sourced from the ILO (International Labour Organization). Employment by occupation group, gender breakdown, informal employment rate, and time-series trend are available. No wage data.

Most countries in Africa, Asia, Latin America, and the Middle East fall here.

Color Layers (what the colors show)

AI Exposure Score

A score from 0 to 10 estimating how much generative AI (large language models, AI assistants, automation tools) will reshape the tasks in this occupation. A higher score does not mean job loss - it means the work will change significantly. Scored independently for each ISCO group.

Score 8–10: Heavy AI integration expected (e.g. Professionals, Clerical workers). Score 1–3: Minimal AI impact (e.g. Elementary physical labour, Skilled trades).

Robotics Risk

A score from 0 to 10 for physical automation exposure - robots, machines, and automated systems replacing manual or physical tasks. This is separate from AI: a factory floor job may score low on AI but high on robotics.

Offshoring Risk

How easily could this work be moved to a different country? High scores indicate tasks that can be done remotely across borders (e.g. software development, data entry). Low scores indicate work that must be local (e.g. construction, healthcare).

WFH Potential

Work From Home potential - how much of this occupation can be performed remotely. Based on task characteristics (digital vs physical, client-facing vs back-office).

Pay Layer

Annual wage in USD for each occupation group, from official government statistics. Tier A countries show this. The color goes from darker (lower pay) to brighter (higher pay). Wages are median unless labeled "mean" - Australia and some EU countries publish averages, not medians.

Growth Layer

Projected 10-year employment growth (%) for each occupation. This data is only available for the United States (from BLS projections) and only shows in the US Detail view. All other countries show "no data" in this layer.

Gender Layer

The female share (%) of workers in each occupation group. Darker color = fewer women. Brighter color = more women. Available for 204 countries via ILO data.

Wages & Pay

Median Wage

The middle salary - half of workers earn more, half earn less. Less distorted by very high earners than an average. Most Tier A countries report median wages. Shown as "median" in the UI.

Mean Wage

The average salary across all workers in the group. Can be pulled higher by a small number of very high earners. Australia and several EU countries only publish mean wages. The UI flags this with a "mean" label in the legend.

OECD Wage Benchmark

The national average annual wage in USD (adjusted for purchasing power) published by the OECD for 38 countries. This is a country-wide number, not occupation-specific. It appears in the insight cards as a reference point - not as a treemap color layer.

Workforce & Economy Terms

Informal Employment

Work that happens outside formal employment arrangements - no written contract, no social protection, not registered with tax authorities. Common in developing economies. WorldJobsData shows the informal employment rate where ILO data is available (144 countries). A high informality rate means the AI-exposure scores should be read with extra caution, since AI adoption patterns in informal economies differ from the formal sector assumptions behind the scores.

India: ~87% informal. Nigeria: ~93% informal. Germany: ~3%.

Female Labor Force Participation (LFP)

The share of women in the total working-age population who are employed or actively looking for work. Shown in the Gender insight card as a country-level figure (not per occupation).

Workforce-Weighted Average

When computing a single headline score for a country (e.g. "average AI exposure"), we weight each occupation by how many people work in it. A large occupation like Professionals has more influence on the headline than a small one like Armed Forces.

15-Year Workforce Shift

The change in employment share for each ISCO group between ~2010 and ~2025, shown in the Trend insight card. It reveals whether a country is shifting toward more knowledge work, more service work, or shedding certain sectors over time.

Getting started

Pick a country

Use the dropdown at the top of the page. 206 countries are available. The treemap loads immediately and shows how the workforce is distributed across occupation groups.

Choose a color layer

Click one of the 7 buttons: AI · Robotics · Offshore · WFH · Pay · Growth · Gender. The treemap recolors instantly. The legend at the bottom left updates to show what the colors mean for that layer.

Hover or click any block

Hovering a block shows a tooltip with the occupation name, employment share, wage (if available), and all four risk scores. Clicking a block in Scatter view highlights it.

Read the insight cards below the chart

Up to 15 cards appear below the visualization for each country. They summarize key findings: workforce composition, AI risk profile, gender breakdown, wage benchmarks, trend over 15 years, and more. Scroll down to see them all.

Views

Treemap

Default view. Each rectangle = one occupation group. Rectangle area = share of total employment. Color = selected layer value. Best for seeing the overall shape of a country's workforce at a glance.

Detail (US only)

Shows all 341 individual US occupations from BLS data in a full-screen treemap. Much more granular - you can see "Software Developers" separately from "Accountants" rather than both bundled into "Professionals". Only available for the United States.

Scatter

AI exposure on the X-axis, median pay on the Y-axis. Each bubble = one occupation group. Bubble size = employment. Instantly shows which groups are both high-paid AND high-AI-exposed - those are the most interesting to watch.

Compare

Click "Compare" to open a two-panel view. Choose two countries and compare their treemaps side by side under the same color layer. Useful for seeing structural differences between economies.

Insight cards - what each one shows

Headline

The workforce-weighted average AI/Robotics/Offshore/WFH score for the whole country. One number that summarises the country's overall exposure on the selected layer.

Risk Profile

Four horizontal bars showing all four dimension scores side by side. Lets you see whether a country's risk is concentrated in AI, Robotics, Offshoring, or WFH.

AI Exposure Mix

A stacked bar breaking down what share of the workforce falls into low / medium / high AI exposure bands.

World Comparison

Where this country ranks among all 206 countries on the selected layer. Shows the top 3 and bottom 3 countries as reference points.

Highest / Lowest Exposure

The top 5 occupation groups with the highest and lowest score on the current layer within this country.

Largest Occupations

The biggest groups by employment share - where most workers actually are.

15-Year Trend

How each ISCO group's share of total employment has changed from ~2010 to ~2025. Growing groups shown in green, shrinking in red.

Gender Mix

Female share per occupation group, plus country-level female labor force participation rate.

Informal Employment

The share of the workforce in informal work. Includes a caveat: high informality means AI-exposure scores should be read with extra caution.

Pay Range

The ratio of the highest-paid to lowest-paid occupation group - a simple measure of wage inequality within the country's workforce.

Pay Rank

How this country's wages compare to a basket of reference countries on a specific occupation (Managers, Professionals, etc.).

OECD Benchmark

The country's national average annual wage vs the OECD average. Available for 38 countries.