Key findings

  • The link is very strong. Across 144 countries the correlation between the share of informal jobs (ILO ILOSTAT) and weighted AI exposure (WorldJobsData) is -0.89. It is the strongest of the country statistics we tested (HDI, schooling, GDP, poverty, unemployment and participation among them).
  • Formal economies score higher. The 37 countries with under 20% informal jobs average 4.90 out of 10. The 48 countries with 80% or more average 3.36.
  • The reason is the job mix. In the most informal group about 40% of workers farm and 13% do office-type jobs. In the most formal group it is 3% and 52%.
  • Informal jobs cover most of the world's workers in our data. In these 144 countries, about 1.36 billion of 2.06 billion workers are in informal jobs (our calculation, using each country's own latest rate).
  • Low exposure is not safety. Informal workers usually have no contract, savings buffer or social insurance. A low AI score only says AI is not the main thing changing their work today.

What "informal work" means

When economists say a job is informal, they mean it has no formal protection. There may be no written contract, no social security, no paid sick leave. Typical examples are a street vendor, a day labourer on a building site, a farmer working family land, or a worker at a small unregistered shop. The International Labour Organization (ILO) measures the informal share of all jobs for most countries. It ranges from about 1% in Switzerland to about 98% in Niger.

AI exposure is our own score from 0 to 10 for how much of a country's work could be done or helped by AI. We weight every occupation group's score by how many people work in it. Higher means more of the country's jobs involve the writing, calculating and information handling that AI tools can already help with.

The pattern: the more informal, the lower the score

Scatter plot of informal employment share against weighted AI exposure for 144 countriesEach dot is a country. Countries with more informal jobs have lower weighted AI exposure, correlation -0.89. Switzerland is near 1% informal with exposure 5.35; Niger is near 98% informal with exposure 3.12. 0% 20% 40% 60% 80% 100% 3 4 5 6 Share of jobs that are informal (%, ILO) Weighted AI exposure (0-10) Switzerland Luxembourg Russia Mexico Nigeria India Niger Brazil Indonesia Under 20% 20% to 39% 60% to 79% 80% and above
Each dot is one country. The dashed line is the straight-line trend (correlation -0.89). Sources: ILO ILOSTAT (informal employment share, latest available year per country) and WorldJobsData weighted AI exposure scores.

The dots fall in a clear diagonal band. Countries such as Switzerland, Norway, Finland and Iceland sit at the top left: almost no informal jobs and exposure near 5. Countries such as Niger, Chad and Madagascar sit at the bottom right: nearly all jobs informal and exposure near 3. Every 10 points of extra informality goes along with about 0.19 fewer points of exposure, on average.

Five bands, one clear ladder

Average AI exposure by share of informal jobsBar chart: under 20% informal averages 4.90; 20 to 39% 4.30; 40 to 59% 3.98; 60 to 79% 3.57; 80% and above 3.36. 0 2 4 6 Under 20% 4.9 20% to 39% 4.3 40% to 59% 3.98 60% to 79% 3.57 80% and above 3.36
Simple average of weighted AI exposure across the countries in each band. Source: ILO ILOSTAT, WorldJobsData.
Informal share of jobsCountriesAverage exposureWorkers coveredFarming shareOffice-type share (ISCO 1-4)
Under 20%374.90327M2.7%52.4%
20% to 39%154.30197M6.3%39.3%
40% to 59%243.98130M12.4%30.6%
60% to 79%203.57273M18.5%21.3%
80% and above483.361,137M39.8%12.5%

The ladder has no gaps: each step toward more informal work lowers the average score. The most informal band is also the biggest by far, with about 1.1 billion workers in our files, and it includes India, Indonesia, Pakistan, Nigeria and Bangladesh. That is why this chart is not just about small or poor countries. It describes where most of the world's workers actually are.

Why informality tracks exposure: farms versus offices

Average share of workers in farming by informality bandShare of workers in farming rises from 2.7% in the most formal countries to 39.8% in the most informal. 0% 10% 20% 30% 40% Under 20% 2.7% 20% to 39% 6.3% 40% to 59% 12.4% 60% to 79% 18.5% 80% and above 39.8%
Average share of workers in farming (ISCO major group 6) in the countries of each band that report all nine major occupation groups.

The bars above show the main driver. In the most informal countries about 40 of every 100 workers farm. In the most formal countries it is about 3 in 100. At the same time, office-type jobs (managers, professionals, technicians and clerks) go from about 13 in 100 up to about 52. Farm work is where our scores are lowest, and office work is where they are highest, so informal economies score lower almost automatically.

That also means informality does not cause a low score. Both come from the same thing: the structure of a country's economy. When a country industrialises, its farm work shrinks, its formal jobs grow, and its AI exposure climbs along the same line.

What it looks like in the biggest economies

CountryInformal share of jobsWeighted AI exposureWorkers covered
Nigeria93.1% (2024)3.3171M
Pakistan88.7% (2025)3.3178M
India87.2% (2025)3.26477M
Bangladesh84.0% (2024)3.2169M
Indonesia81.0% (2023)3.44139M
Vietnam67.0% (2024)3.2754M
Mexico56.9% (2025)3.8359M
Brazil35.6% (2025)4.13102M
Russia19.2% (2025)4.6273M

India has the largest workforce in this list and one of the most informal (87.2%), with exposure of 3.26. Russia is the most formal of the large economies (19.2% informal) with exposure of 4.62. Brazil (35.6%) and Mexico (56.9%) sit in the middle. Rates are the latest available year for each country, so they are not all from the same year.

Countries that break the pattern

We drew the trend line and measured how far each country sits above or below it. We only used countries that report all nine major occupation groups, so that a missing group cannot distort the average.

More exposed than their informality predicts

CountryInformal shareActual exposureExpected from informalityDifference
Nauru93.9%4.493.22+1.27
Luxembourg6.0%5.874.87+1.00
Barbados61.8%4.493.82+0.67
Netherlands5.4%5.444.88+0.56
United Kingdom19.8%5.084.61+0.47

Less exposed than their informality predicts

CountryInformal shareActual exposureExpected from informalityDifference
Grenada48.0%3.244.08-0.84
Tunisia36.9%3.614.29-0.68
Romania2.4%4.264.93-0.67
Ethiopia77.7%2.943.53-0.59
Bahamas18.2%4.154.64-0.49

The largest gaps are about 1.3 points on the 10-point scale, but they are rare. 137 of the 139 countries in this check land within one point of the line, and the typical country sits 0.17 points from it.

What this means for workers

There is a common misreading of numbers like these: "informal workers are safe from AI." Our scores do not say that. A street vendor in Niamey is not in a job AI can do, but the vendor also has no unemployment insurance, no pension and no employer who can retrain them. If AI changes the economy around them, for example by changing what customers buy or what factories hire for, they have little cushion.

For countries with large informal workforces, the more realistic AI question is not "which jobs vanish" but "who benefits from tools like translation, voice assistants and phone-based services." In formal economies the question is the opposite: which office tasks change first, and how fast can workers move.

Limits to keep in mind

See any country's informality and AI exposure

Open a country page to see its job mix, its informal share and its occupation-level AI scores.

Open the explorer →

Was this analysis useful?

Methodology and sources. Informal employment share is the ILO ILOSTAT indicator stored in each country file (informal_employment_rate, latest year available, 2009 to 2025). AI exposure is each country's weighted average exposure (stats.weighted_avg_ai_exposure), from occupation scores weighted by employment (ILO ILOSTAT and national statistics offices). We compared the 144 countries that have both. Correlation is Pearson's r. Band averages are simple averages of countries. Farming and office-type shares use the countries that report all nine ISCO major groups (139 of 144). Billions of informal workers multiply each country's rate by its total employment, so they mix years. All of these are our calculations.

AI exposure scores are model-estimated: an AI language model (Claude Opus 4.7) scored each ISCO-08 occupation group on 28 May 2026, guided by published research. They are informed judgements, not measurements, and not forecasts of job losses. Read the full methodology and how the scores were checked.

Frequently asked questions

What is informal employment?
Informal employment means jobs without formal protection, such as no written contract, social security or paid leave. The ILO counts street vendors, casual labourers, many farm workers and unregistered business owners in this group.
Do countries with more informal jobs have lower AI exposure?
Yes, by our scores. Across 144 countries the correlation between informal employment share and weighted AI exposure is -0.89. Countries under 20% informal average 4.90; those at 80% or more average 3.36.
Does low AI exposure mean informal workers are safe?
No. Exposure only measures how much of the work AI could help with. Informal workers often have no savings buffer, no social insurance and no contract, so any change to their income hits harder.
Why are formal economies more exposed to AI?
They have more office-based work, which AI tools handle well. In countries under 20% informal about 52% of workers do office-type jobs. In countries at 80% or more informal it is about 13%.

Related analyses

AI risk and poverty: the double burden Why farm workers are the most AI-proof Richer countries face more AI exposure (HDI) How fast will AI disruption arrive in each country? AI job risk country ranking 2026 United States: which jobs are most at risk from AI in 2026 United Kingdom: clerical workers most exposed US vs World: how America's AI job risk compares globally India country page

Sources

  • ILO (2026) - ILOSTAT, informal employment rate by country, latest available year (CC BY 4.0)
  • ILO and national statistics offices - employment by occupation, latest available year per country (CC BY 4.0)
  • WorldJobsData (2026) - Occupation-level AI exposure scores, methodology

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