Women Face Twice the AI Automation Risk — ILO Data From 138 Countries Reveals the Uncomfortable Truth

The Numbers That Should Change the Conversation

When we talk about AI replacing jobs, the conversation usually stays abstract — "some jobs will be automated, some won't." But a massive new dataset from the International Labour Organization cuts through the vagueness with uncomfortable precision. After analyzing 2,861 distinct work tasks with input from 1,640 subject-matter experts across 138 countries, the ILO has produced the most granular picture yet of who generative AI is actually coming for. [Fact] ILO Working Paper 140

The headline finding: 1 in 4 workers globally now faces meaningful exposure to generative AI. [Fact] ILO Working Paper 140 That is not a prediction about the future. It is a measurement of the present.

But the real story is not the average. It is who sits inside that average.

The Gender Gap Nobody Planned For

The ILO's March 2026 research brief lays it out plainly: occupations dominated by women — business administration, clerical support, customer service — face 29% exposure to generative AI. Male-dominated occupations — construction, manufacturing, transport — face just 16%. [Fact] ILO Research Brief, March 2026

That is nearly twice the exposure rate. And the gap gets worse when you look specifically at automation risk — the subset of exposure where AI does not just assist workers but potentially replaces the task entirely.

For female-dominated occupations, the automation risk stands at 16%. For male-dominated occupations: 3%. [Fact] ILO Research Brief, March 2026 That is more than a five-to-one ratio on pure displacement risk.

Let that sink in. The occupations where women are concentrated are not just more exposed to AI — they are more than five times as likely to see tasks fully automated rather than augmented.

The Root Cause: Occupational Segregation

A companion ILO publication released the same day, analyzing 84 countries, names the structural driver behind these numbers: occupational segregation. [Fact] ILO, "Gen AI, Occupational Segregation and Gender Equality"

In 88% of countries analyzed, women face higher workplace risks from generative AI than men. [Fact] ILO News, March 2026 That is not a regional anomaly. It is a near-universal pattern driven by which occupations women have historically been channeled into.

The geographic picture is striking. In Switzerland, the United Kingdom, the Philippines, and several Caribbean and Pacific island developing states, more than 40% of women's employment is exposed to generative AI. [Fact] ILO News, March 2026 High-income countries overall see 41% of jobs exposed, compared to just 11% in low-income countries. [Fact]

As ILO co-author Anam Butt puts it: "Discriminatory social norms shape who enters which occupations," resulting in women concentrated in the very roles most vulnerable to automation. [Fact — direct quote] ILO News, March 2026

And the pipeline to fix this is thin. Women comprised only 30% of the AI workforce in 2022 — an increase of just 4 percentage points since 2016. [Fact] ILO News, March 2026 Women are overrepresented in the roles AI threatens and underrepresented in the roles building AI.

Who Is in the Crosshairs?

The ILO's Working Paper 140 identifies 3.3% of global employment sitting in the highest exposure category. [Fact] ILO Working Paper 140 That sounds small until you break it down by gender: 4.7% of female workers fall into that highest-risk bracket, compared to 2.4% of male workers. [Fact] ILO Working Paper 140

Women are nearly twice as likely as men to be in the most exposed group globally. In high-income countries, the disparity is even starker: 9.6% of women versus 3.5% of men. [Fact] ILO Working Paper 140

The occupations driving this pattern are ones you can probably name without looking them up. Administrative assistants — scheduling, correspondence, document management — these are textbook generative AI tasks. Secretaries managing communications and documentation. Receptionists handling inquiries and routing information. Bookkeeping clerks processing transactions and reconciling records. Data entry keyers transferring information between systems.

These are not obscure job categories. They employ tens of millions of people worldwide, and they are overwhelmingly filled by women.

The Income Divide Makes It Worse

Here is where the ILO data gets genuinely concerning. The overall exposure rate varies enormously by country income level: just 11% in low-income countries versus 34% in high-income countries. [Fact] ILO Working Paper 140

At first glance, that might seem like good news for developing nations — less exposure means less disruption. But the ILO's companion report, "Disruption Without Dividend," argues the opposite. Workers in low-income settings who could benefit from AI augmentation — using AI tools to become more productive rather than being replaced — often lack the digital infrastructure to access those tools. [Fact] ILO "Disruption Without Dividend"

Meanwhile, the workers who are vulnerable to automation in those same countries face displacement without the safety nets — retraining programs, unemployment insurance, social protection — that higher-income countries can deploy. [Claim — ILO analysis] The result is a lose-lose: miss the productivity gains, absorb the displacement costs.

ILO senior economist Janine Berg frames the stakes clearly: "With the right policies, we can avert reinforcing existing discrimination." [Fact — direct quote] ILO News, March 2026

For registered nurses in developing countries, this dynamic is especially relevant. Nursing involves a mix of tasks — documentation, care planning, patient communication — where AI augmentation could genuinely improve outcomes. But only if the infrastructure exists to deploy it.

What This Means for Workers in Exposed Roles

If you work in one of the highly exposed occupations, the ILO data suggests three things worth knowing.

First, exposure does not equal replacement. The ILO deliberately distinguishes between exposure (AI can perform some of your tasks) and automation risk (AI can replace them entirely). Many exposed workers will see their roles change, not disappear. Administrative assistants who currently spend 60% of their time on scheduling and correspondence may find those tasks automated — but the remaining 40% involving judgment, coordination, and relationship management becomes more valuable, not less.

Second, the transition speed varies by country. If you are in a high-income country, the changes are happening faster because the infrastructure and investment are already there. In lower-income settings, the same changes may take years longer to materialize, but they are coming.

Third, the gender dimension demands policy attention. This is not a natural disaster — it is a pattern created by which tasks AI happens to be good at first. Text generation, data processing, information synthesis — these capabilities landed first, and they happen to overlap heavily with female-dominated clerical and administrative work. [Claim — editorial analysis] Construction work, plumbing, and electrical installation are not inherently safer from AI; they are simply not yet within reach of current AI capabilities. That could change.

Explore how AI affects these roles: Administrative Assistants, Secretaries, Receptionists, Customer Service Representatives, Bookkeeping Clerks, Data Entry Keyers, Registered Nurses.

Sources

1. ILO Working Paper 140, "Generative AI and Jobs: A Refined Global Index of Occupational Exposure," May 2025 (updated methodology). Link
2. ILO Research Brief, "GenAI and the Gender Gap," March 2026. Link
3. ILO, "Disruption Without Dividend: How Digital Divide and Task Differences Split the Impact of AI," March 2026. Link
4. ILO, "Gen AI, Occupational Segregation and Gender Equality in the World of Work," March 2026. Link
5. ILO News, "New ILO Data Confirm Women Face Higher Workplace Risks From Generative AI Than Men," March 2026. Link

Update History

• 2026-03-21: Added ILO occupational segregation report data — 88% of countries finding, 40%+ exposure in Switzerland/UK/Philippines, women at 30% of AI workforce, ILO official quotes, added secretaries and customer service representatives to occupation links
• 2026-03-19: Initial publication based on ILO Working Paper 140 and March 2026 research briefs

---

This article was researched and written with AI assistance using Claude (Anthropic). All statistics are sourced from ILO publications as cited. This is AI-generated analysis of publicly available international labor research and should not be taken as professional career or employment advice. We encourage readers to consult the original ILO sources linked above for full methodology and findings.