AI could force 11 million Americans to change jobs by 2035. Here’s who is most exposed

Artificial intelligence may not eliminate work on the scale of some of the more dramatic predictions surrounding the technology, but it could radically change who does it, how it is done, and how many people are required to do it.

A new analysis from the McKinsey Global Institute estimates that around 11 million Americans, equivalent to roughly 7% of the current US workforce, may need to switch occupations by 2035 as AI reshapes the economy. That would mean around 770,000 occupational transitions every year, compared with a historical average of about 215,000. McKinsey says the pace of change could therefore be three to four times faster than previous technological shifts.

The important word is “occupations”. The forecast does not mean 11 million people will necessarily become unemployed. It means millions could find that the work they currently do is shrinking, changing substantially, or becoming less economically valuable.

Which jobs face the greatest risk?

The biggest exposure is concentrated in work built around repetitive, predictable, rules-based tasks.

Office and administrative work is particularly vulnerable. Receptionists, administrative assistants, general office clerks, bookkeeping and accounting clerks, data-entry workers, scheduling staff, and similar roles increasingly involve tasks that software can perform quickly and cheaply.

McKinsey’s earlier research estimated that demand for clerical roles could decline by 1.6 million jobs by 2030, alongside potential declines of 830,000 retail salesperson positions, 710,000 administrative assistant positions, and 630,000 cashier positions.

Customer service is another major exposure zone. Chatbots and AI agents can already handle a growing range of routine enquiries, troubleshooting, booking, billing, and information requests. The more predictable the interaction, the easier it is to automate.

Retail faces a similar combination of automation and digitisation. Cashiers, ticket clerks, inventory processing, basic merchandising, and parts of customer assistance can increasingly be handled through self-service systems, computer vision, AI agents, and automated inventory platforms.

Logistics and warehousing will also change, although the impact is more complicated. AI can optimise routing, inventory, scheduling, demand forecasting, warehouse allocation, and dispatch. Robotics can increasingly handle physical movement in controlled environments. Human workers will remain essential in many unpredictable physical settings, but fewer people may be required for some routine logistics tasks.

Data-heavy professional work is another emerging area of exposure. Computer programming, market research, marketing, sales support, financial analysis, and medical-record processing all contain significant quantities of tasks that AI can assist with or potentially automate. An Anthropic analysis, for example, identified computer programmers, customer-service representatives, data-entry workers, medical-record specialists, market research and marketing professionals, and sales representatives among occupations with high AI exposure.

The harder problem: moving people between occupations

The economic challenge is therefore less about replacing a single task and more about moving millions of people from shrinking occupations into expanding ones.

McKinsey’s earlier work found that workers in lower-wage occupations can face substantially higher probabilities of needing to change occupations than higher earners. Brookings has also found that around 6.1 million highly AI-exposed US workers, primarily in clerical and administrative jobs, have relatively limited capacity to absorb a major career transition because of factors including limited savings, age, local job opportunities, and narrow skill sets.

That makes retraining a much bigger issue than simply teaching people how to use ChatGPT. The World Economic Forum estimates that 39% of workers’ existing skill sets could be transformed or become outdated between 2025 and 2030. At the same time, employers expect analytical thinking, resilience, flexibility, creative thinking, technological literacy, and AI and big-data skills to become increasingly important.

How workers can build career resilience

The first strategy is become an AI user rather than an AI-dependent worker. Knowing how to use AI for research, analysis, drafting, coding, customer service, data processing, or workflow automation can increase the value of an existing professional skill set.

Second, move up the value chain. Routine execution is easier to automate than judgement. An accountant who only prepares standard reports faces a different risk from one who interprets financial performance and advises management. A marketer who only produces copy faces a different risk from one who understands customers, positioning, strategy, and brand economics.

Third, develop domain expertise alongside technology skills. AI literacy without subject knowledge can quickly become commoditised. Combining the two creates a more defensible position.

Fourth, strengthen human skills that remain difficult to automate. Leadership, negotiation, relationship-building, persuasion, empathy, complex judgement, mentoring, and handling ambiguous situations are increasingly valuable precisely because they sit outside purely repetitive workflows.

Finally, treat learning as an ongoing part of the job. A once-in-a-decade qualification may be less useful than a habit of continuously adding relevant skills.

The broader labour-market picture is also more complicated than a simple jobs-versus-AI equation. The World Economic Forum expects significant growth in technology, healthcare, education, construction, delivery, and several other frontline roles by 2030, even as clerical and secretarial occupations decline.

That suggests the defining career skill of the AI era may be adaptability itself. The safest job may not be one that AI cannot perform. It may be a career in which the worker can repeatedly move towards the parts of the job where humans continue to create the most value.