About half of all American jobs could be replaced by artificial intelligence and robotics, according to a new analysis from the McKinsey Global Institute, which found that currently available technologies are capable of automating more than half of all U.S. work hours.
The report — Agents, Robots and Us — concludes that if companies fully redesigned their workflows to take advantage of AI systems, both cognitive and physical tasks could be rapidly automated across wide sectors of the economy. Roughly 40% of the jobs most at risk involve drafting, processing information and routine reasoning — tasks that generative AI agents can already perform. The impact is beginning to be felt: hiring has slowed for paralegals, administrative staff, office support roles and even programmers, McKinsey said.
Physical and hazardous jobs, including warehouse roles or machine operation, are also highly vulnerable to replacement by robots.
However, the consultancy found that around one-third of U.S. jobs are difficult to automate because they rely on attributes machines cannot replicate, such as empathy, dexterity, care and physical presence. About 70% of tasks performed by carers and healthcare workers fall into this category. Building maintenance and repair work — requiring judgment, adaptability and situational awareness — was also deemed unlikely to be automated.
McKinsey noted that widespread adoption of AI will depend largely on policy decisions and investment levels rather than technical limitations. Redesigning workflows, not simply automating isolated tasks, could add an estimated $2.9 trillion a year to the U.S. economy by 2030.
The research suggests human labour will continue to play a critical role, with workers shifting into hybrid roles where people and AI collaborate. Teachers, for example, may use AI to generate content or handle administrative work, while retaining responsibility for interpretation, judgment and decision-making. Across the workforce, the report predicts people will spend less time preparing documents and more time supervising and directing AI systems.
It also highlights the emergence of new job categories — from AI product managers to safety specialists — as artificial intelligence still requires human oversight, tuning and testing.
Businesses are already beginning to restructure around AI. Fintech firm Klarna, which listed in New York in September, has said the technology is enabling it to expand revenue without increasing headcount. Law firm Clifford Chance and telecoms giant BT have recently announced reductions in business services roles as they integrate more AI into daily operations.
Evidence of AI’s uneven impact is mounting. A recent Stanford report found that early-career workers aged 22–25 in AI-exposed occupations have experienced a 13% decline in employment, while more experienced employees have been largely unaffected.
The McKinsey study follows a Microsoft report that identified translators, sales representatives and financial advisers as among the first professions likely to feel the brunt of AI disruption. By contrast, nurses, ship engineers, plasterers and water treatment operators were seen as among the safest from automation.
