AI threatens millions of back-office jobs, women face the biggest risk

The first big AI shock for middle-class workers may not come from software engineers. It is already aimed at the back office, where human resources, payroll, billing and other clerical tasks sit inside jobs that are heavily concentrated among women.

Brookings estimates that about 6.1 million U.S. workers are in jobs with high AI exposure and low adaptive capacity, and 86% of those workers are women. The vulnerable roles are clustered in clerical and administrative work, including secretaries, payroll and timekeeping clerks, and court, municipal and license clerks, exactly the kind of work that keeps offices, schools, hospitals and local governments running.

The scale of that risk matters because office and administrative support occupations employed 16,856,793 people in the United States in 2024, according to U.S. labor data, and women made up 73.4% of that workforce. That makes the coming AI disruption more than a technology story. It is a labor-market story about where routine white-collar work is concentrated, who performs it, and how quickly employers can automate tasks once handled by people.

The International Labour Organization says women are disproportionately exposed to generative AI because they are overrepresented in jobs most susceptible to automation and underrepresented in AI-related and STEM occupations. Its 2025 update says generative AI is more likely to change tasks, skills and working conditions than to wipe out entire occupations overnight, but that should not be mistaken for reassurance. If AI tools absorb scheduling, form processing, document review, payroll reconciliation and other repetitive office work, the pressure lands first on the jobs built around those functions.

That is why the policy response matters now. The ILO warns that choices made today, especially on social dialogue, occupational segregation and women’s representation in AI roles, will help determine whether the technology expands decent work or widens existing gaps. For millions of workers in the administrative economy, the threat is not a distant automation wave. It is a near-term restructuring of the tasks that have long supported the middle class.