AI job losses to hit women clerical workers hardest: study

WASHINGTON, UNITED STATES — A new analysis reveals a stark divide: there is a group of United States workers are most at risk of AI-driven job displacement. 

Research published in January 2026 by the Brookings Institution and researchers from the Center for the Governance of AI finds that while many highly exposed workers are resilient, a vulnerable group of 6.1 million—overwhelmingly women in clerical roles—lack the capacity to adapt.

“AI supplanting human workers in these jobs would continue a decades-long pattern of advancements in information technology handling work formerly done by people,” Mark Muro, a Senior Fellow at the nonpartisan Brookings Institution, told CBS News.

Adaptive capacity defines AI job loss risk

This capacity is a composite index based on four occupation-level factors: 

• Net liquid wealth
• Growth-weighted skill transferability
• Geographic density
• Age

The analysis links this new metric with established AI exposure scores to create a more complete picture of potential workforce vulnerability.

The results indicate that the relationship is positive and very strong: the greater the AI exposure in an occupation, the higher its adaptive capacity. 

For example, software developers, financial managers, lawyers, and other highly exposed professions are often associated with high salaries, financial cushions, broad skill sets, and strong professional networks. 

Therefore, 26.5 million out of 37.1 million employees in the four deciles with the highest AI exposure also have above-median adaptive capacity, placing them in an advantageous position to switch jobs if they are displaced.

However, Ben May, Director of Global Macro Research at Oxford Economics, notes, “We’re skeptical that firms can quickly and seamlessly substitute workers with AI even in sectors where the potential for AI disruption is greatest.”

Women in clerical jobs face high AI risk, low resilience

Despite the overall trend, a significant vulnerable cohort exists. Approximately 6.1 million workers face the double burden of high AI exposure and low adaptive capacity. 

These employees are clustered in clerical and administrative jobs, including general office clerks (2.5 million total employees), secretaries and administrative assistants (1.7 million), and receptionists and information clerks (965,000), where the savings will be small, and the transferability of skills is low.

More importantly, these vulnerable workers are mostly women (86%). This difference underscores the fact that the displacement caused by AI can strongly affect female-dominated fields of work. 

The report includes such high-risk jobs as court clerks (85% female), secretaries (96% female), and medical secretaries (94% female), which has high exposure scores and very low adaptive capacity scores as measured, which means that the risk of losing a substantial portion of income and struggling to find a new job is very high in case of displacement.

As Muro notes, “It’s a bit more about what women do in the economy rather than what they are.”

U.S. regions where AI job disruption will hit hardest 

The distribution of high-risk workers in the United States is uneven. Whereas tech centers such as San Jose, California, and Seattle are not only dense in highly exposed and highly adaptive workers, vulnerability is concentrated elsewhere. 

The proportion of employees in high-exposure, low-adaptive-capacity jobs is high in small metropolitan regions and college towns, especially in the Mountain West and Midwest.

The most concentrated areas are those that house high levels of workers in administrative and clerical positions. The main destination places are college towns such as Laramie, Wyoming; Huntsville, Texas; and Stillwater, Oklahoma. 

Also state capitals such as Springfield, Illinois, Carson City, Nevada, and Frankfort, Kentucky, and towns in New Mexico and Oklahoma.

This vulnerable group of metro-area workers accounts for an average of 3.9% nationally and as much as 6.9% in certain regions, indicating that some regional labor markets might be particularly vulnerable to the disruptive impact of AI.

The analysis notes that “as AI continues to spread across the economy, adaptability analysis can provide a starting point for policymakers to better understand who may be most in need of better support to weather job transitions.”

The study reveals that the future of work may be shaped not merely by which jobs AI can perform, but by a profound societal divide between those with the financial and professional resilience to adapt and a vulnerable, predominantly female clerical workforce for whom this technological shift presents a double jeopardy of high exposure and low capacity to transition.