The history of technological progress offers a blunt warning for the AI era: when machines replace routine cognitive work, women have often been the first to feel the impact. That pattern is now re-emerging as artificial intelligence moves from experimental tools to mainstream workplace systems, threatening a broad range of clerical, administrative and customer-service roles where women remain heavily represented.
Automation's Uneven Burden
The concern is not that AI will eliminate only women's jobs, but that it is more likely to erode occupations in which women are disproportionately employed. Economists have long observed that automation tends to target tasks rather than entire professions, and the tasks most vulnerable to software-driven substitution are often repetitive, structured and information-based. Those characteristics describe a large share of office support work, scheduling, transcription, basic bookkeeping, call handling and other functions that have historically employed more women than men.
The historical analogy is instructive. When telephone switchboards were automated, the work of connecting calls — once a major source of employment for women — largely disappeared. Later, the spread of personal computers transformed office work again, reducing demand for many clerical tasks while increasing the premium on technical and managerial skills. In both cases, the technology did not affect all workers equally. It reshaped the labor market in ways that often hit women harder because of the occupational segregation already built into the economy.
AI is different in scale, but not necessarily in direction. Generative models can draft emails, summarize documents, answer routine queries, process forms and assist with scheduling at a speed and cost that make them attractive to employers seeking efficiency gains. That means the first wave of disruption may land in jobs that are not only administrative, but also deeply embedded in the service economy — a sector where women make up a large share of the workforce in many countries.
Lessons From Past Shifts
The broader lesson from earlier waves of automation is that technology rarely distributes gains and losses evenly. Workers with more access to training, higher levels of digital literacy and greater mobility into technical roles tend to benefit first. Those in lower-paid, routine-heavy positions often face the most immediate displacement. Because women are still overrepresented in many of those roles, the AI transition risks widening existing labor-market gaps unless policy and corporate strategy respond quickly.
That risk is amplified by the way AI is being adopted. Companies are not necessarily replacing entire departments overnight; instead, they are using AI to compress workloads, reduce headcount growth and redesign workflows. In practice, that can mean fewer entry-level positions, fewer support staff and less room for workers to move up through traditional career ladders. For women, who are already more likely in some economies to face career interruptions related to caregiving, the loss of these stepping-stone jobs could have long-term consequences for earnings and advancement.
The issue is not limited to white-collar work. AI-enabled tools are also moving into retail, hospitality, healthcare administration and other service sectors where women are heavily represented. While some of these technologies may improve productivity and create new roles, the transition period is likely to be uneven. Employers may capture the efficiency gains quickly, while workers absorb the uncertainty, retraining costs and job churn more slowly.
Policy Response Needed
That makes the policy challenge urgent. Governments and central banks are already watching AI for its implications for productivity, inflation and labor-market slack, but the distributional effects deserve equal attention. If AI raises output while concentrating job losses among women, the result could be a more unequal labor market even in a stronger economy.
The response will need to go beyond generic calls for reskilling. Training programs must be aligned with the jobs AI is creating, not just the jobs it is destroying. Employers will also need to rethink hiring pipelines, promotion structures and workplace flexibility so that workers displaced from routine roles can move into higher-value positions. Without that, the AI boom could repeat a familiar pattern: broad technological progress paired with concentrated harm for those already closest to the automation frontier.
The central economic question is no longer whether AI will change work. It will. The more pressing question is who will pay the price for that change. History suggests women may be among the first to do so unless policymakers and businesses act with unusual speed and precision.