🔍 Read the full analysis: AI’s Place In A Long History Of Workplace Change on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A Nov. 21, 2026, article from ThorstenMeyerAI.com compares AI with earlier workplace technologies through the histories of elevator operators, typists and typesetters. It reports that automation usually changes tasks rather than eliminating entire occupations, while jobs centered on one task can decline sharply; the article also cautions that the 1945 New York elevator strike is not proven to have caused the shift to automatic elevators.
ThorstenMeyerAI.com published a historical report on Nov. 21 examining how automation affected elevator operators, typists and typesetters, and what those changes may suggest about AI’s place in workplace history. The account argues that technology more often takes over tasks than eliminates whole occupations, but that jobs built around a narrow set of tasks can shrink quickly.
The report draws on economist James Bessen’s study of 271 occupations in the 1950 U.S. census, tracked through 2010. Bessen found that, among the occupations he examined, only elevator operator could be said to have largely disappeared because of automation. Other occupations vanished for varied reasons, including changing demand and technologies that made earlier services obsolete. The report uses the finding to distinguish automation of work tasks from the disappearance of an entire job category.
It also tracks the rise and fall of typing work. The U.S. census counted 3.9 million stenographers, typists and secretaries in 1970. In 1988, the Bureau of Labor Statistics counted 985,000 word processors and typists; the report says the occupation fell by 51% by 2000, far more than the 6% decline official forecasters had expected. By 2025, it counted about 40,400. Those figures concern the occupational category used in the cited data, not every worker who performs typing as part of another job.
The third example is typesetting. The account says German print workers, mainly typesetters, went on a three-week strike in 1978 that slowed changes to how text design work was organized. In 1998, the typesetter apprenticeship was merged into a newer media-design trade. The examples show different paths: some roles contract or change, while workers and occupational structures can adapt rather than disappear outright.
Infographic 01 · The number to remember
The one job automation wiped out
Elevator operators rose with the skyscrapers and fell with the automatic elevator. Of 271 occupations in the 1950 census, it is the only one that automation largely wiped out.
US elevator operators · census · thousands
Story vs. evidence
- Documented (TIME, 8 Oct 1945): 16,600 elevator operators and building workers on strike in 2,100 Manhattan buildings for five and a half days.
- Documented (Otis, company account): 1948 a recorded voice to help passengers trust automatic elevators; 1950 the first building with Otis high-speed, no-operator elevators (Dallas).
- Not documented: that the strike made owners switch to automatic elevators. A popular story; no source we found shows it.
Census figures: economically active persons 14 and over, each census in its own classification (1950 in the 1950 classification). Today the BLS publishes no separate count of elevator operators; installers and repairers are a different occupation.
Sources: US Census Bureau, Historical Statistics of the United States (1975), series D 233-682 (row D 587); J. Bessen, How Computer Automation Affects Occupations (2016); TIME, 8 Oct 1945; Otis company history; BLS Employment Projections 2025-35 (27 Aug 2026).
Powered by Thorsten Meyer AI
What Earlier Job Losses Show
The history offers a useful distinction for readers trying to judge claims about AI and employment: automating tasks is not the same as automating a whole occupation. An office worker’s duties may change as software handles drafting, sorting or routine processing, without eliminating the entire role. The report’s historical examples caution against treating a technical capability as proof that a job will disappear.
At the same time, broad reassurance can obscure concentrated losses. The report argues that where a job is made up largely of one task, a tool that performs that task can put the occupation under sharper pressure. It cites an August 2026 BLS projection that office and administrative support would be the fastest-shrinking occupational group through 2035, partly because of AI tools. About one in nine U.S. jobs is in that group, according to the report. The projection is a forecast, not a count of jobs already lost or proof that AI alone will cause the decline.
For workers, employers and policymakers, the distinction affects how change is measured and addressed. Occupational totals can miss substantial shifts in duties, skills and working conditions inside jobs that remain. Conversely, evidence that most historical automation was partial does not guarantee that future AI will follow the same pattern or spread at the same pace.
Products Worth Considering

Vari ComfortEdge 48×30 Inch Adjustable Electric Standing Desk

Dell UltraSharp U4025QW 40-Inch Curved 5K2K WUHD LED Monitor

FLEXISPOT 36-Inch Height-Adjustable Standing Desk Converter with Keyboard Tray
As an affiliate, we earn on qualifying purchases.
From Lift Attendants to Office Work
Elevator operation grew with the rise of multistory buildings. U.S. census counts cited by the report rose from about 13,000 operators in 1900 to roughly 97,000 in 1950. Automatic elevators existed before they became widely accepted, and Otis’s company history says it offered recorded voices in 1948 to help passengers adjust to unattended systems. In 1950, an Otis Autotronic installation without attendants opened in Dallas’s Atlantic Refining Building, according to the company history.
The report also addresses a frequently repeated account of the transition: that a 1945 New York elevator strike prompted building owners to replace operators. TIME magazine’s Oct. 8, 1945, coverage reported that 16,600 elevator operators and other building service workers struck at 2,100 buildings for five and a half days over a wage ruling. Union leader David Sullivan said, “We were hurting our best friends—the garment workers.” The report says it found no document establishing that the strike caused the later switch to automatic elevators, so the proposed link should be treated as a popular story, not a demonstrated fact.
The source places these examples in a longer sequence of workplace technologies, from typewriters and word processors to personal computers and AI tools. Its focus is not a claim that history repeats exactly, but what previous changes reveal about the difference between a tool taking over a task and an occupation disappearing.
““In only one case—elevator operators—can the decline and disappearance of an occupation be largely attributed to automation.””
— James Bessen, economist, as quoted in the report
Limits of the Historical Comparison
The historical record does not establish that the 1945 strike caused building owners to adopt automatic elevators. The report says it found no document proving that causal link, despite its recurrence in later accounts. It also notes that later retellings sometimes use strike totals larger than the figure TIME reported at the time.
The BLS projection cited for office and administrative support is a forecast through 2035, and the source says AI tools are only part of the explanation. It does not provide a breakdown of how many projected job changes would result from AI, other technologies, economic conditions or changes in demand. The report also does not establish how quickly AI will affect different occupations, which tasks will change, or whether new roles will offset reductions.
Historical counts are not perfectly interchangeable: categories, job titles and census classifications change over time. The examples can clarify past patterns, but they cannot by themselves predict the scale or distribution of AI’s future effects.
Watch Jobs and Tasks Separately
The report’s next step is comparative rather than a prediction: it places the three occupations alongside earlier episodes in its series on automation. For readers following AI’s workplace effects, the relevant evidence will include updated employment data, changes in occupational duties and BLS projections, as well as whether affected workers move into new roles.
The cited BLS outlook runs through 2035. As that period approaches, future releases can show whether office and administrative support employment tracks the projection and how much change is attributed to AI tools. Until then, the historical cases support a distinction, not a guarantee: AI may change tasks within jobs, while some occupations may face much steeper declines.
Key Questions
What is the main finding of the report?
The report says automation has historically changed tasks more often than it eliminated whole occupations. It cites James Bessen’s analysis of 271 occupations from the 1950 U.S. census through 2010, in which elevator operator was the one occupation whose disappearance could largely be attributed to automation.
Did the 1945 elevator strike cause operators to be replaced?
That causal link is not established. The report says the strike is documented, but it found no document showing that the walkout caused building owners to switch to automatic elevators.
How much did U.S. typing work decline?
The report says the BLS counted 985,000 word processors and typists in 1988 and about 40,400 in 2025. It reports a 51% decline by 2000, compared with a 6% decline expected by official forecasters. These figures refer to the named occupational category.
Does the history prove AI will eliminate fewer jobs?
No. The report uses earlier automation as context, not as a forecast. It says historical patterns show that machines often take over tasks, but also that jobs centered on a narrow set of tasks can decline sharply. AI’s future effects and pace remain uncertain.
Source: ThorstenMeyerAI.com
