As artificial intelligence (AI) replaces jobs, women may be hit harder than men, according to an analysis. Because AI is concentrated on repetitive, standardized clerical and administrative work, job categories where women make up the majority are directly exposed to risk. The argument cites the past automation of telephone switchboards and the spread of personal computers (PCs), when female workers stood at the front line of employment shocks.
According to research published by the International Labour Organization (ILO) in March of this year, about one in 10 women working in high-income countries is employed in an occupation with the highest risk of job displacement by AI. Ariane Hegewisch, a senior researcher at the Institute for Women’s Policy Research in the United States, noted that “AI is focused on repetitive tasks, number processing, and finding patterns,” adding that “these are exactly the clerical and administrative jobs where many women work.” She added that “90% of secretaries and office workers are still women.”
There are already precedents of technology shocks affecting one gender asymmetrically. In the 1940s, about 95% of American telephone operators were women. For white women born in the United States under age 20, it was the occupation with the largest share of employment. The job required months of training, and during peak call times, operators had to accurately connect hundreds of calls per hour—a decently paid job at the time.
The rupture began when AT&T introduced automated switchboards. Considered one of the largest automation investments in U.S. history, the decision led to the loss of about 75,000 telephone operator jobs over several decades. In a 2020 study, Boston University professor James Feigenbaum and Duke University professor Daniel Gross analyzed that the introduction of automated switchboards stripped existing operators of their jobs and careers. Many were pushed into lower-paying jobs or dropped out of the labor market entirely.
David Autor, a professor at the Massachusetts Institute of Technology (MIT), explained that “most people do not change careers often,” and “when the value of the occupation they chose disappears due to technological advancement, they generally end up doing worse work than before.”
The next generation, however, adapted differently. According to the same study, younger women found jobs at similar levels in other fields in step with the changing labor market. The concern at the time that future generations would have fewer job opportunities largely proved unfounded.
There are also cases where technology actually expanded women’s economic activity. Household appliances such as washing machines, which spread in the 1950s, significantly reduced the time women spent on unpaid housework, widening their room to enter paid labor. From 1950 to 1980, the number of employed women in the United States increased by more than 25 million, and the female labor force participation rate rose from 34% to 52%. Over the same period, the number of employed men increased by 16 million, but their participation rate fell by 5 percentage points.
Women’s entry into the labor market created new demand for services. Childcare and food preparation that had been done at home shifted to paid services, and jobs increased in retail, dining, leisure, and hospitality. As companies grew in size, demand for clerical work such as payroll management and customer account processing also expanded.
The spread of PCs in the 1980s was another shock. In the 1980 U.S. Census, about 4.7 million Americans worked as secretaries, stenographers, and typists, and more than 98% of them were women. With IBM’s PC launch in 1981 and the arrival of Microsoft’s word processor “Word” in 1983, executives could write reports, book business trips, and print documents themselves, and some roles of existing office workers disappeared.
But PCs also opened new opportunities. Women expanded their scope of work by using tools such as spreadsheets and databases. Technology functioned as a means of replacing existing tasks while also raising productivity. A 2022 study by Marcus Dillender and Eliza Forsythe found that after companies adopted software for clerical and administrative work, employment in that sector fell 1%, but wages for college-educated workers rose 3%.
Similar changes could emerge in the AI era. Rather than eliminating entire jobs, AI may take over repetitive tasks and help workers focus on work that requires human judgment.
The fact that today’s women are more educated at the college and graduate level than previous generations and hold a higher share of managerial positions is cited as a factor that could partially mitigate the risk of job losses from automation. However, as AI’s potential to replace even high-skilled work is discussed, it is uncertain how long such protective effects will last. “AI can learn,” warned senior researcher Hegewisch.
For market participants, this analysis is another clue to gauging the profit structure of the AI investment boom. This is because how automation reshapes employment and wages in specific job categories can change the landscape of consumer spending and service demand. What past cases showed was not the disappearance of jobs themselves but a shift in the composition of tasks. The speed at which AI encroaches on clerical and administrative areas, and the capacity to absorb the new roles it creates, are cited as variables that will determine the direction of the job market going forward.