Industries that are expected to shrink the most because of automation are food services, customer service and sales, and office support. Women are overrepresented in these sectors — and hold more low-paying jobs than men — so they stand to be more affected, the report finds.
Black and Hispanic workers, workers without college degrees, and the youngest and oldest workers also are more likely to have to find new jobs by 2030, the study says.
According to the report, by 2030, at least 12 million workers will need to change jobs as the industries in which they work to shrink — 25 percent more than the institute predicted in a report published in February 2021. Most of those workers will be at the low end of the pay scale, and they probably will need to acquire new skills before they can transition into new industries.
The report says the labor market also will be upended over the next decade by the government’s investments in green technology, the growing demand for health-care workers as the U.S. population ages and the structural changes to the workforce brought about by the pandemic. It says these trends will converge with advancements in artificial intelligence to increase demand for some existing jobs, create new jobs for new industries and make other jobs obsolete.
Today’s low-wage workers are the most vulnerable to job losses by 2030 across all categories, according to the McKinsey report. It finds that workers earning less than $38,200 could account for almost 80 percent of all potential career transitions in that period. This means that retail salespeople, cashiers and other low-wage workers — among whom a larger proportion are women — are particularly vulnerable.
Although advances in artificial intelligence will make some jobs obsolete, it also could have some positive effects on existing jobs and create new work opportunities, according to the report. For white-collar workers, automation could mean less time doing rote or technical tasks, and more time spent on creative or strategic work that artificial intelligence cannot do — yet. The report finds that lawyers and civil engineers are among the workers who stand to benefit most. But workers in more-manual fields, such as health care or agriculture, do tasks that cannot be automated as easily.
“We see generative AI enhancing the way STEM, creative, and business and legal professionals work rather than eliminating a significant number of jobs outright,” the authors wrote.
But these fields are male-dominated. According to the U.S. Bureau of Labor Statistics, in 2022, women accounted for only 17.1 percent of civil engineers and 38.5 percent of lawyers.
While new technologies are expected to create new jobs, those jobs may not all be desirable, says Kerry McInerney, a research fellow at the Leverhulme Center for the Future of Intelligence at Cambridge University. Workers who are more likely to be in low-paying jobs with long hours or difficult conditions today could in the future “get pushed into areas like data labeling,” which is the process of adding labels to videos, images or audio that teach machine learning models to recognize what is in them. Those jobs “can be psychologically very harmful,” because of the nature of the material that has to be identified, McInerney says.
The findings align with existing research showing that women will be affected by the waves of workforce automation differently from men.
An analysis of Goldman Sachs data published in April by Mark McNeilly, a marketing professor at the University of North Carolina at Chapel Hill’s Kenan-Flagler Business School, and Paige Smith, an MBA candidate at the school, found that 8 in 10 female workers in the United States, compared to 6 in 10 men, have jobs that are “highly exposed” to automation, meaning that over a quarter of their tasks can be automated by generative AI.
The report suggests that training and retraining workers in the skills of the future will be a major challenge for employers. It also could be an opportunity, they argue, to “recruit from populations that are often overlooked,” such as older workers, workers without college degrees, workers with disabilities or employment gaps, and those who have been incarcerated.
Employers also could use AI to find and hire these kinds of candidates, the report suggests.
But research on the current use of artificial intelligence in hiring suggests that AI “doesn’t grapple very well with different kinds of life experiences, different patterns of coming in to work,” McInerney says. It may ask a candidate who has just had a baby what she does in her spare time, for instance, and rank that candidate on the basis of her answer without taking into account that the candidate may not have as much time for hobbies, she says.
Automated hiring systems that are based on AI may be able to find candidates with nontraditional backgrounds, she says, “but they’re not necessarily going to treat them equitably.”