AI May Push 11 Million US Workers Into New Careers by 2035

Some 11 million US workers may need to change careers entirely by 2035 because of artificial intelligence, according to projections reported by WJRT ABC12. That figure is not a forecast of unemployment. It is a forecast of occupational displacement — people whose current jobs shrink enough that staying in the same field stops being viable.

The 2035 horizon matters. This is not next quarter's jobs report; it is a decade-long structural shift comparable in scale to the long decline of US manufacturing employment, but compressed into a wider set of white-collar occupations.

The national picture

At the national level, the projection suggests a labor market in which career changes become more common and arrive with less warning. A worker who today plans a 20-year arc in a single occupation may instead face two or three forced pivots. The 11 million figure implies roughly one in fifteen US workers, given a labor force of about 170 million, will need a genuine career change rather than a routine job switch within the same field.

Where the pressure concentrates

Displacement will not spread evenly. Roles built around routine cognitive work — data entry, basic document processing, scheduling, entry-level analysis — face the earliest and sharpest exposure. Workers in these roles compete not with other workers but with software that performs the task at near-zero marginal cost.

Roles involving physical presence, manual dexterity, licensing requirements, or sustained human judgment face lower projected exposure. That split, rather than education level alone, increasingly determines who needs a career change and who does not.

What this means for an individual worker

Translate the aggregate number into personal terms. If your day consists primarily of tasks a system can complete from a text prompt, the projection implies your occupation's headcount will shrink over the next decade — and the earlier you reposition, the less crowded the transition will be. Concretely, that means auditing your weekly tasks, identifying which ones are routine and repeatable, and building toward work that combines domain knowledge with judgment, client relationships, or physical execution.

For workers in lower-exposure occupations, the projection cuts differently: as routine roles contract, demand for people who can supervise, verify, and integrate AI outputs is likely to grow. Positions adjacent to automation have historically paid a premium over the positions automation replaced.

Read the numbers with care

Projections at this distance carry real uncertainty. The 11 million figure depends on assumptions about adoption speed, regulation, and business behavior that will shift over the next decade. Long-range labor forecasts have historically overstated some displacements and understated others — the offshoring projections of the early 2000s, for example, missed the occupations that technology created entirely. Treat the number as a direction of travel, not a schedule.

What is less uncertain is the asymmetry: the cost of preparing for displacement that never arrives is small, while the cost of ignoring displacement that does arrive falls entirely on the unprepared worker. WJRT ABC12's report points to a decade in which career mobility stops being an occasional event and becomes a standing requirement of US employment.