Artificial intelligence and automation could push roughly 11 million American workers into entirely different occupations by 2035, even as the technologies help create more jobs than they eliminate, according to a new McKinsey Global Institute report.

The report, released Tuesday, estimates that about 7% of the current U.S. workforce could need to change occupations over the next decade as businesses adopt AI and other forms of automation.

McKinsey's projections vary significantly depending on how quickly companies adopt the technology and how much automation reduces demand for human labor. The number of workers needing to switch occupations could range from roughly 6 million to more than 16 million.

Under McKinsey's base scenario, automation could reduce labor demand by the equivalent of approximately 36 million jobs by 2035. At the same time, economic growth, demographic changes, investment and expansion of the AI economy could generate roughly 41 million jobs.

That would leave the economy with more jobs created than eliminated, but millions of workers could still find themselves needing to develop new skills or move into completely different occupations.

McKinsey estimates that roughly 25 million workers affected by automation could remain in their existing occupations because economic growth would offset declining demand. Another 11 million, however, would need to make a more significant career transition.

“The next decade’s challenge is mobility, not scarcity,” the report stated.

Making those transitions could be difficult. McKinsey found that only about one in seven workers who need to change occupations would have a relatively direct path into a growing field requiring little retraining while maintaining their pay.

Approximately 41% would face what the report describes as a “winding” path involving significant retraining, while another 45% would face an “unpaved” path involving larger skill gaps, credential requirements, or other barriers.

The changes will not necessarily mean that millions of people suddenly lose their jobs. McKinsey estimates that more than 70% of workers could see significant changes to their existing jobs as technology takes over or reallocates at least 15% of the time currently spent on individual tasks.

The workers most likely to require occupational changes are concentrated in several major sectors. More than three-quarters of those potentially needing to switch occupations currently work in office and administrative support, retail and sales, or transportation and logistics.

Customer service representatives, retail sales associates, office assistants, cashiers and warehouse workers together account for roughly one-third of the projected occupational transitions.

At the same time, the report identifies several areas expected to see substantial employment growth, including health care support, health care professionals, construction and management. Those categories account for more than one-third of the positions expected to emerge in growing occupations.

The pace of workforce movement could also increase dramatically. McKinsey estimates that roughly 770,000 workers would need to move between broad occupational groups every year through 2035, about 3.6 times the historical annual average of approximately 215,000.

Employers are already changing the skills they seek from workers. The report found that demand for “AI fluency” in job postings increased roughly elevenfold between 2022 and 2026, while demand for adaptability increased fivefold.

The report comes as AI becomes an increasingly prominent economic and political issue in the United States, with candidates across the country discussing artificial intelligence, automation and the expansion of data centers as part of their economic platforms.

McKinsey ultimately describes the coming transformation less as a story of disappearing work than one of large-scale workforce reallocation. The challenge will be helping workers move into the growing occupations as technology continues changing how Americans work.