AI in Business
AI and Your Job: What the Government's New Ten-Year Forecast Really Says
The Bureau of Labor Statistics now rates every occupation for its exposure to AI. The most exposed jobs are still projected to grow. The ones expected to shrink tell a narrower, more useful story.
For three years the question has hung over every office: will AI take my job? In August the federal government gave its most detailed answer yet. Alongside its new ten-year employment projections, the Bureau of Labor Statistics did something it had never done before. It rated 831 occupations for how exposed they are to artificial intelligence.
The headline inside those ratings is not the one most people expect.
By the numbers
A new way to measure the risk
BLS projects the economy will add 5.9 million jobs between 2025 and 2035, growth of 3.5 percent. That is slower than the decade just ended, and it is the backdrop for everything else.
The new piece is a set of AI exposure categories: low, moderate, high and very high. Each occupation is placed by comparing what AI could do with its tasks against how people are actually using AI at work. BLS is careful about what the label means. In its words, exposure "does not imply job loss, productivity gains, automation probability, or wage effects."
The surprise inside the numbers
We added up the projections by exposure group using the BLS data file. The most exposed group, 206 occupations holding 31 percent of all jobs, is projected to grow 2.0 percent by 2035. That is slower than the 3.5 percent average, but it is growth, not collapse. Only 60 of those 206 occupations are projected to shrink, and 87 are projected to outgrow the economy as a whole.
Some of the most exposed jobs are among the most promising. Software developers rate very high on exposure and are still projected to add 174,700 jobs, a gain of 10.2 percent. Management analysts are projected to grow 10.1 percent, accountants and auditors 5.0 percent, and lawyers 4.7 percent. Nurse practitioners, rated high, are the fastest-growing occupation in the entire projection, up 41 percent.
A job can be full of tasks AI touches and still grow, because the work that remains becomes more valuable. The jobs most at risk are the ones where the routine tasks are the whole job.
Where the losses land
Those jobs are concentrated in one place. Office and administrative support is projected to lose 752,100 jobs, a drop of 4.0 percent and the largest decline of any major occupational group. BLS names the cause directly: automation tools, "including those powered by AI," are likely to reduce demand for several of these occupations. By our count from the BLS file, about 88 percent of jobs in the group sit in occupations rated very high for exposure.
The individual lines are familiar. General office clerks are projected to lose 156,200 jobs. Customer service representatives are projected to lose 141,800, and the agency's own handbook notes that more of their tasks continue to be automated. Secretaries and administrative assistants lose 114,100, bookkeeping and accounting clerks 85,600, and data entry keyers a quarter of their ranks.
BLS also expects generative AI to limit demand for some jobs in arts, design and media, and AI tools in the sales process to keep trimming sales occupations, which are projected to shrink 1.4 percent overall.
Where AI is creating work
The same projections credit AI with new demand. Data scientists are projected to grow 34.6 percent, adding 95,400 jobs. Professional, scientific and technical services are projected to add 926,700 jobs, driven in part by demand for AI systems, research and consulting. Computing infrastructure and web hosting grows 25.1 percent. Even utilities grow, partly on the electricity AI consumes.
For the self-employed and creative
Readers who work for themselves should look closely at their own line. Writers and authors, nearly two-thirds of whom are self-employed, rate very high on exposure and are projected to be roughly flat, down 0.3 percent. Graphic designers are projected to dip 1.7 percent and photographers 0.7 percent. Computer programmers, a separate occupation from software developers, are projected to fall 7.3 percent. These projections count wage and salary jobs, so they describe the market a freelancer competes in more than freelancing itself.
What businesses are actually doing
Adoption is real and rising. In the Census Bureau's latest survey, 23.8 percent of businesses used AI in some business function in the two weeks ending September 20, up from 17.7 percent in January.
So far, very few say it has cost anyone a job. In a Census working paper covering late 2025 and early 2026, 95.7 percent of AI-using firms reported no change in headcount from AI over the previous six months. About 2 percent reported a decrease and 2.3 percent an increase. Two-thirds of users said they use AI only to help with tasks people still do.
Read the fine print
BLS is candid about the limits. Its projections assume technology advances at roughly its historical pace, and the agency says uncertainty about AI's effects "remains very high." The exposure ratings reflect AI as it stood around mid-2023, mostly language models, and two of the five sources behind them are usage data from AI companies, Anthropic and Microsoft, mapped to job tasks. The working paper has not been through the review Census gives its official publications. And no projection is a forecast of any one person's career: BLS advises reading them for direction and relative size, not precise values.
What it means for your career
- Look at your tasks, not your title. The jobs projected to shrink are made mostly of routine processing: entering, filing, scheduling, answering the same question again.
- Move toward judgment. The exposed jobs that grow, from software to nursing to analysis, pair AI-touched tasks with decisions people still want a person to make.
- Learn the tools early. With nearly a quarter of businesses already using AI, and most using it to help staff rather than replace them, fluency is quickly becoming part of the job description.
Get the Report
Modern Enterprise Report follows AI where it meets real work and real income: the data, the tools, and the opportunities they open. Subscribe free.
Sources
- Bureau of Labor Statistics, Employment Projections, 2025 to 2035 (news release, August 27, 2026)
- Bureau of Labor Statistics, AI exposure categories for occupations
- Bureau of Labor Statistics, Occupational Outlook Handbook: Customer Service Representatives
- U.S. Census Bureau, Business Trends and Outlook Survey, national estimates
- U.S. Census Bureau, Center for Economic Studies, Working Paper CES 26-25, AI use and employment at U.S. firms
A note on how we work. Modern Enterprise Report is free to read. We may be paid by partner companies whose offers appear on this site, never by you. Nothing here is investment advice or a promise of income, and results vary with the opportunity, the market and the effort you put in.
