AI is likely to hit female-dominated occupations harder than other parts of the US labor market.
This unpleasant news comes from a new study from the Brookings Institution and the Centre for the Governance of AI.
The report finds that many of the roles most exposed to automation are clerical and administrative jobs that have long employed disproportionately high numbers of women. While AI is expected to reshape nearly every sector of the economy, researchers say the burden of displacement could fall unevenly, exacerbating existing gender and economic inequalities.
Clerical roles face elevated risk
Drawing on public labor data and proprietary analysis from Lightcast, a labor market analytics firm, the researchers examined how susceptible different occupations are to AI-driven automation and how easily workers in those roles could transition into new jobs.
They found that more than six million workers face a particularly difficult path if their jobs are eliminated by AI, due to a combination of factors such as age, limited savings, and fewer transferable skills. According to the study, 86% of those workers are women.
Many of the most exposed roles include secretaries and administrative assistants, payroll and timekeeping clerks, and court or municipal clerks — occupations that rely heavily on routine tasks such as scheduling, data entry, and document processing. These functions are increasingly being handled by AI-powered software that can perform them faster and at lower cost.
Mark Muro, a senior fellow at the nonpartisan Brookings Institution, said the threat posed by AI is part of a longer technological trend rather than a sudden shock.
“These are occupations that have been under attack for a long time,” he told CBS News.
Decades of decline
Muro noted that advances in information technology — from spreadsheets to enterprise software — have steadily reduced the need for human labor in administrative work. AI represents an acceleration of that process, not a complete departure from it.
Importantly, he stressed that the findings should not be interpreted as a reflection of workers’ abilities.
“It’s a bit more about what women do in the economy rather than what they are,” he said.
Historically, women have been concentrated in support roles that were viewed as stable entry points into the middle class, particularly for workers without advanced degrees. As those jobs disappear or shrink, the loss could ripple outward, affecting household income, retirement security, and local economies.
Researchers warn that without targeted retraining or policy intervention, displaced workers may struggle to find roles that offer similar pay or stability.
Not all are equally vulnerable
While the report paints a stark picture for some workers, it also highlights a more optimistic reality for others. Roughly 70% of employees in AI-exposed jobs are likely to transition into new positions with comparable earnings, the researchers said.
Those workers tend to be in fields such as marketing, finance, science, and technology, where jobs require broader analytical, creative, or managerial skills. These roles are often augmented rather than replaced by AI, allowing workers to shift responsibilities instead of losing employment entirely.
Sam Manning, a senior research fellow at the Centre for the Governance of AI, said adaptability is closely tied to skill diversity.
“There’s a wider range of things that you need to leverage in [these jobs], versus some of these more administrative, back office-type roles where the core competencies are more narrowly defined,” he told CBS News.
In these higher-skilled occupations, AI may increase productivity or change workflows, but it is less likely to fully eliminate the human role.
So far, so limited
Despite growing concern about AI-driven job losses, economists say the technology has not yet led to widespread displacement across the US economy.
In a report released earlier this month, Ben May, director of global macro research at Oxford Economics, said there is little evidence that companies are currently replacing large numbers of workers with AI. He also questioned whether AI will significantly raise unemployment in the near future.
“We’re skeptical that firms can quickly and seamlessly substitute workers with AI even in sectors where the potential for AI disruption is greatest,” he said.
This may come as a surprise to some in the tech sector, as stories of job cuts tend to dominate the headlines. This week, Amazon confirmed it has laid off 16,000 employees after an internal email mistakenly revealed the job cuts.
Experts say adoption costs, regulatory uncertainty, and the complexity of real-world jobs have slowed the pace of automation. Still, they caution that the impact could grow as AI systems become more capable and affordable.
How workers can prepare
While policymakers debate how to respond, researchers say workers can take proactive steps to reduce their vulnerability.
Manning and Muro encourage employees to engage with AI tools rather than avoid them, learning how the technology can enhance their productivity or expand their skill sets.
“Trying to experiment with this technology to see how you can make yourself more productive and expand your capability set to stand out within whatever field you’re in is definitely something that can help,” Manning said.
The study’s authors argue that retraining programs, employer investment, and public policy will be critical to ensuring that AI does not widen existing gender gaps in the workforce.
A new film premiering at Sundance explores growing anxiety around AI, examining its power and what it could mean for the future of humanity.
