UN study finds women are 3x more likely to face AI job disruption

When people debate which jobs AI will eliminate first, the conversation usually revolves around software engineers, graphic designers, or copywriters. But a new study suggests the hardest-hit group might be sitting in plain sight, doing the everyday office work that keeps companies running.

A joint study from the United Nations' International Labour Organization (ILO) and Poland's National Research Institute (NASK) reveals that jobs traditionally held by women in high-income countries are nearly three times more likely to be disrupted by generative AI than those held by men.

Specifically, the researchers found that 9.6% of roles held by women are poised for transformation, compared to just 3.5% of men's jobs.

Clerical work is directly in the crosshairs

Overall, the report estimates that 34% of all jobs in high-income nations will feel the impact of generative AI. While software and finance positions are among the earliest to evolve, administrative and clerical positions face the highest vulnerability.

That disparity comes down to occupational sorting. According to the US Census Bureau, women held roughly 95% of secretary and administrative assistant jobs between 2000 and 2019. The US Department of Labor also ranks administrative support as the fifth most common occupation for women.

Summarizing meeting notes, managing calendars, processing receipts, and sorting emails are exactly what current large language models do best. The study emphasized that exposure does not mean these roles vanish overnight, but rather that a significant portion of daily duties will be handed off to software.

The dangerous adoption gap

Turning AI from a threat into an assistant sounds straightforward on paper, but actual workplace adoption tells a different story.

Harvard Business School professor Rembrand Koning pointed out that while women could use generative AI to augment and protect their jobs, they are currently far less likely to actively experiment with the technology than their male peers.

That creates a compounding risk: the very roles facing the steepest automation pressure are also the ones where workers are trailing behind in adopting the tools that could save them.

Why it matters

For workers in administrative, scheduling, or operational support roles, this data is an urgent wake-up call. The real question is not whether AI will change your job, but who will direct the tool. Waiting around for corporate training is a mistake; professionals in clerical roles should immediately start testing generative AI tools for routine tasks like draft generation, data cleanup, and inbox management.

For businesses, simply viewing AI as an easy excuse to cut administrative headcount is a trap. Administrative staff manage countless unwritten organizational nuances that raw models cannot handle. We think companies that use AI to upskill their existing clerical workforce will see far better productivity gains than those that aggressively cut support staff and leave remaining teams drowning in logistics.

Ultimately, AI has no gender bias, but the labor market does. When automation targets administrative tasks, systemic imbalances make women bear the initial brunt. Fixing that requires narrowing the adoption gap right now before automation becomes full replacement.