Artificial intelligence could deepen racial and geographic gaps in U.S. economic mobility, with Black and Latino workers facing higher projected displacement shares than White and Asian workers by 2035, McKinsey & Company said.
The finding comes from “Advancing US economic mobility in the age of AI,” published Nov. 19 by the McKinsey Institute for Economic Mobility. The research examined how artificial intelligence may affect work and opportunity among Black, Latino and rural populations.
The authors drew on McKinsey Global Institute research, U.S. Bureau of Labor Statistics data and American Community Survey data. They mapped work activities across about 900 occupations and compared projected changes in labor demand between 2025 and 2035.
Forty percent of automation-exposed hours worked by Black employees and 41% of those worked by Latino employees could be displaced rather than augmented. The comparable shares were 37% for White employees and 36% for Asian employees, the authors found.
McKinsey defines displacement as automation that reduces employers’ need for labor. Augmentation refers to automation that instead supports workers through higher productivity, new activities or other mechanisms that absorb freed capacity.
The differences stem partly from the kinds of tasks workers perform. Standardized activities with easily checked results and limited consequences for mistakes are more likely to be displaced, while work requiring judgment in higher-stakes settings is more likely to be augmented.
Physical activities accounted for 35% of Black workers’ hours and 39% of Latino workers’ hours, compared with 26% for White workers, according to the analysis. Physical tasks made up 34% of rural workers’ hours, against 28% for urban workers.
Robotics and other artificial intelligence systems could make physical work more exposed to automation, although capital and infrastructure needs may slow adoption compared with software-based systems. Other groups are more exposed to generative artificial intelligence, with adoption more often augmenting work than replacing it.
About 21% of current U.S. work hours, much of it routine work in entry-level jobs, could be reduced by automation by 2035 across technologies. That estimate comes from a separate McKinsey Global Institute report published Sept. 29, 2026.
Automation adoption could be similar across racial groups and urban and rural areas, while its effects differ because workers are concentrated in different occupations. Black workers exceed their workforce share in each of the six occupations with the steepest projected declines in labor demand.
Demand for postal mail sorters and processors is projected to fall by more than 50% between 2025 and 2035. Black workers hold 37% of those jobs, about three times their workforce share, the authors said. Latino workers account for 43% of hand packers and packagers, while rural workers hold 28% of postal mail carrier and dishwasher jobs.
The United States has also seen a long-term decline in upward mobility. According to Opportunity Insights data cited in the report, the share of children earning more than their parents fell from about 90% for the 1940 birth cohort to about 50% for children born in the 1980s.


