Robots and employment
Will automation wipe out factory jobs? BLS projects a 1% decline
US production jobs are projected to decline by just 1.0% from 2023 to 2033, while total US employment grows 4.0% to 174.6 million. The government's own forecast does not show automation wiping out factory work.
That sits awkwardly against the widely quoted estimates that tens of percent of jobs are at risk. The Bureau of Labor Statistics, whose whole job is projecting employment, forecasts production occupations nearly flat over the decade.
This page traces the projection to its source and sets the official forecast beside the automation-risk headlines.
Data covers US employment projections, 2023 to 2033 (BLS). Published 2026-07-15. Last reviewed 2026-08-16. Last updated 2026-08-16. Edited by Mike Ramsey / Reliable Media.
The figures and where they come from
Each figure is rated for how safely you can cite it today. Ratings judge current usability, not whether a number was ever correct.
| Figure | What it is | Source | Citation Confidence | Notes |
|---|---|---|---|---|
| -1.0% (2023-33) | Projected production-jobs change | [A] | High | Production occupations are projected to decline just 1.0% over the decade, not collapse. The official forecast is near-flat. |
| +4.0% | Projected total employment growth | [A] | Medium | Total US employment is projected to grow 4.0% over the same period, so production jobs decline only slightly against a growing whole. |
| 174.6 million | Projected total employment, 2033 | [A] | Medium | The projected 2033 total, up about 6.8 million jobs over ten years. |
Why the numbers disagree
The automation-risk headlines and the BLS projection are answering different questions, and only one is a forecast. Studies that put tens of percent of jobs at risk estimate technical susceptibility to automation. The BLS projection is an actual employment forecast that already weighs technology, demand, demographics, and cost, and it puts production jobs down just 1.0% over the decade.
That gap is instructive. A job can be technically automatable and still exist in ten years, because automation is gated by cost, integration, regulation, and whether firms choose to do it. The BLS projection reflects those frictions; the risk studies deliberately ignore them to measure susceptibility.
None of this means automation has no effect. Production employment declining while total employment grows 4.0% is consistent with automation slowly shifting the mix of work. But it is a slow reshaping in the official forecast, not the sudden collapse the risk headlines imply.
How to cite these figures
Cite the projection as the official forecast: US production jobs down 1.0% from 2023 to 2033, with total employment up 4.0% to 174.6 million.
Contrast it with the risk estimates, and explain the difference: a projection weighs cost and adoption; a susceptibility estimate does not.
Avoid presenting either as the whole truth. Susceptibility is high on paper; the forecast decline is small. Both are real answers to different questions.
Where people go wrong
Treating an automation-risk percentage as a job-loss forecast. The BLS forecast for production jobs is a 1.0% decline, far smaller.
Reading a small projected decline as no effect. Production employment falling while the total grows is consistent with a slow shift in the mix of work.
Quoting the projection without its horizon. It is a 2023-to-2033 forecast, not a statement about today.
How we checked
The figures come from the BLS 2023-2033 Employment Projections. We cite a state workforce agency's report of them for a narrower reason than blocking: the BLS projection pages fetch normally, but they publish these numbers in linked tables rather than in the page text, so a text-matching citation check has nothing to anchor against. The South Carolina page states the figures in prose, which is what let us fetch and confirm the 1.0% production decline, the 4.0% total growth, and the 174.6 million total.
We set the official forecast against the automation-risk estimates deliberately, because the contrast is the point: a forecast that weighs adoption frictions lands far below a susceptibility estimate that ignores them.
We note this is a projection with a stated horizon, not a description of the present, and we do not claim it settles the automation debate; it is the official forecast, one input among several.
Full source list
Primary sources, with live links. Every figure above traces to one of these.
- [A]SC Department of Employment and Workforce (reporting BLS projections)December 2024
US Bureau of Labor Statistics 2023-2033 Employment Projections, as reported by the South Carolina Department of Employment and Workforce
https://dew.sc.gov/labor-market-information-blog/2024-12/us-employment-projections-2023-2033
Common questions
- Will automation destroy factory jobs?
- The official forecast does not show it. The BLS projects US production jobs to decline just 1.0% from 2023 to 2033, while total employment grows 4.0%. That is a slow shift, not a collapse.
- Why is this so different from '47% of jobs at risk'?
- They answer different questions. Risk studies estimate technical susceptibility to automation; the BLS projection is an actual forecast that already weighs cost, demand, and whether firms adopt. Susceptibility ignores those frictions.
- Does a 1% decline mean automation has no effect?
- No. Production jobs declining while total employment grows 4.0% is consistent with automation slowly shifting the mix of work. It is a gradual reshaping in the forecast, not a sudden loss.
- Whose projection is this?
- The US Bureau of Labor Statistics 2023-2033 Employment Projections. We cite a state workforce agency's report of them because BLS publishes these projections in downloadable tables rather than in page text, so our citation check cannot confirm them on the BLS page itself.
More data, traced to source
- '47% of jobs at risk': what Frey and Osborne actually measured
The famous claim that nearly half of jobs will be automated traces to a single Oxford study. It estimated the probability that occupations are susceptible to computerisation rather than predicting that 47% of jobs will vanish. Here is what it really said.
- 9% of workers, 23% of hours, 47% of jobs? GAO says nobody knows
Estimates of what automation threatens range from 9% to 47%, and they do not share a denominator: one counts workers, one counts work hours, one counts jobs. The US Government Accountability Office reviewed them and concluded the underlying data is not good enough to say. The disagreement is the finding.
- US manufacturing jobs: 17.1 million in 2000, 12.6 million by 2018
US manufacturing employed 17.1 million people at the start of the century and 12.6 million by the time Congress's research service last counted, in 2018. Here are the figures behind 'manufacturing is dying' and 'manufacturing is back,' traced to that report.
- Who keeps the robots running, and what they earn: $64,520
The people who keep factory robots and machinery running are industrial machinery mechanics: 439,600 of them, a median of $64,520 a year, in a field growing much faster than average. A different, larger job than 'robotics technician.'
- Jobs at risk from automation, by country: the OECD puts it at 14%
The famous 47% figure has a direct rebuttal. Applying a task-based method to survey data across 32 countries, the OECD put the share of highly automatable jobs at 14%, ranging from 6% in Norway to 33% in Slovakia.
- Robotics engineer salary: what the $122,930 figure actually measures
The $122,930 median quoted for robotics engineers is O*NET's wage for Engineers, All Other, the residual category robotics engineering sits inside. Here is what it measures, and how it compares with the technicians and mechanics who maintain the same machines.
- Two-fifths of your skills go stale in five years, and that is an improvement
Employers expect 39% of a worker's existing skills to be transformed or outdated within five years. The figure is quoted as evidence that change is accelerating, but it has fallen from 44% and from a pandemic peak of 57%.