McKinsey research says artificial intelligence and automation will not simply erase work. By 2035, the U.S. economy is expected to have more jobs than today, even as the mix of occupations changes.
The catch is scale. Roughly 11 million Americans may need to leave a shrinking occupation for an entirely different one, and an average of 770,000 workers a year will need to cross occupational lines over the next decade. McKinsey says that is 3.6 times the historical rate, and unlike the COVID-era disruption, it would last at least a decade.
The biggest losses are expected in office and administrative support. McKinsey says automation could take on 80 percent of the hours worked in that occupational group, which it describes as the one facing the largest job losses by 2035. An administrative assistant, the researchers say, could move into project management because many of the core tasks overlap.
That shift would not be frictionless. Many project-manager postings require a bachelor’s degree and a professional certification, even when the skills fit. McKinsey says the transition offers a substantial wage gain, but also demands large investments in schooling and credentialing.
Other transitions are even tougher. A dishwasher moving into home health aide work would face a short training requirement but would still need to learn patient assistance and vital-sign assessment. McKinsey says many employers also prefer a nursing assistant credential that takes longer. It adds that 92 percent of healthcare support jobs are in the bottom two wage quintiles.
A packager moving into production fabrication would face a different set of obstacles. McKinsey says about 52 percent of the worker’s skills would transfer, and the move would bring a meaningful raise. But the new role would still require training on specific equipment and advanced quality-control procedures.
Credentials are the most common barrier. McKinsey says credentials are required in 85 percent of growing jobs. About 38 percent are mandated by law, while 47 percent reflect employer preferences for a degree, certificate, or job title. In McKinsey’s view, nearly half of the credential problem is a hiring convention rather than a safety rule.
The report also says even short training programs can be hard to afford because of forgone wages. A worker with modest savings or a child to care for may decline a pathway that pays off over 10 years if it first requires months without pay.
McKinsey says employers, states, colleges, worker-training agencies, and governments all have a role in making transitions easier. It recommends hiring on demonstrated skills, dropping degree screens not required by law, pruning occupational licensing requirements, and creating modular credentials, apprenticeships, and retraining incentives.
The market has already voted on the direction of travel: AI is not just a productivity story, it is a labor-market sorting machine. The private sector will reward workers who can move, retrain, and prove value fast. The question is whether the bureaucratic credentialing system will adapt before millions of Americans are forced to pay the bill in lost time, lost wages, and lost opportunity.
