🔍 Read the full analysis: What Do Vanished Jobs Tell Us About AI Today? on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
A November 21, 2026 report from ThorstenMeyerAI.com examines elevator operators, typists and typesetters to assess what past automation can tell us about AI and work. The history suggests that automation more often changes tasks than eliminates whole occupations, though jobs centered on a narrow set of tasks can decline sharply.
A November 21 report from ThorstenMeyerAI.com uses the histories of elevator operators, typists and typesetters to examine what earlier automation may—and may not—say about AI’s effect on jobs. The report’s central finding, drawing on economist James Bessen’s study of U.S. occupations from 1950 to 2010, is that automation usually took over tasks rather than entire occupations; jobs built around a narrow set of tasks could still contract sharply.
The report cites Bessen’s analysis of 271 detailed occupations tracked from the 1950 U.S. census through 2010. Bessen concluded that only one occupation’s decline and disappearance in that group could be largely attributed to automation: the elevator operator. The report stresses that this is a historical finding, not a forecast of how AI will affect today’s workforce.
Typists offer a different example: a large occupation that fell dramatically as word processors and personal computers spread. The source says the United States had 985,000 word processors and typists in 1988, and that the occupation had about 40,400 workers by 2025. It also reports that official forecasters had expected a 6% decline by 2000, while the actual decline was 51%. These figures describe the occupation’s change over time; they do not by themselves establish how much of the loss was caused by any one technology.
The report’s third case, typesetting, illustrates how workers may contest and shape technological change. It describes a three-week strike in Germany in 1978 over control of text design and says the typesetter apprenticeship was merged into a media-design trade in 1998. The source also points to an August 2026 Bureau of Labor Statistics projection that office and administrative support work would shrink faster than other broad occupational groups by 2035, partly because of AI tools.
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 Past Job Loss Can Tell Us
The cases matter because public discussion of AI often treats job automation as a simple choice between jobs disappearing and jobs staying unchanged. The historical record presented here points to a more mixed process: technology can remove particular duties, alter skills and reduce demand for a role without eliminating an occupation outright. That distinction matters to workers, employers and policymakers trying to understand which kinds of work may be exposed.
At the same time, the typist example cautions against treating the rarity of fully automated occupations as reassurance that few workers can be affected. When a role consists largely of tasks a new technology can perform, demand for that occupation can fall steeply. The report’s history helps frame questions about AI, but it cannot establish the scale or pace of future job losses.
Popular right now

Vitamix 5-Speed Immersion Blender, 625-Watt, Stainless Steel/Black

Dyson V9 Motorbar Cordless Vacuum

Ninja Foodi 10-Quart DualZone 6-in-1 Air Fryer, Black, DZ302
As an affiliate, we earn on qualifying purchases.
From Elevator Cars to Office Software
Elevator operators were once a substantial occupation as American cities built upward. The report gives census counts rising from about 13,000 in 1900 to roughly 97,000 in 1950, before falling to about 77,000 in 1960 and 37,000 in 1970. Automatic elevators existed before operators disappeared, but passenger acceptance took time. Otis’s company history says the firm offered recorded voices in 1948 to help passengers adjust to automatic systems while attendants remained in elevators. In 1950, an unattended Autotronic system opened in a Dallas office building.
The report also addresses a frequently repeated account of a 1945 Manhattan strike prompting building owners to replace elevator operators. TIME reported at the time that 16,600 elevator and building-service workers struck in 2,100 buildings for five and a half days over a wage ruling. The source says it found no document showing that the strike caused the later shift to automatic elevators, so it treats that connection as a popular story rather than an established cause.
These examples differ from the AI changes now under way. Elevators replaced a particular function, while office software changed how typing and document production were done. The report uses them as historical comparisons, not as evidence that AI will reproduce the same timelines or outcomes.
The Limits of the Historical Comparison
The report does not provide a detailed forecast of how many current jobs AI will eliminate, how quickly changes may occur, or which workers will be affected most. The Bessen study covers 1950 to 2010, before today’s generative AI systems, so its findings cannot settle the consequences of a different technology.
The source also cautions that the link between the 1945 strike and the adoption of automatic elevators is unproven. Its account gives occupational counts and a BLS projection, but those figures alone do not isolate AI’s contribution from other economic and technological forces. The cited 2035 projection is a forecast, not a record of job losses already observed.
Watching Employment and Work Tasks
The next evidence will come from employment data and from closer tracking of how AI changes particular duties inside occupations. The report cites the BLS’s August 2026 projection for office and administrative support through 2035; that forecast can be compared with later employment figures as they become available. Changes in job titles, duties, training and hiring will also help show whether AI is reshaping work or reducing demand for specific roles.
For now, the historical cases support a limited conclusion: some occupations have vanished or contracted after technological change, while automation has more often altered parts of jobs. The scale and distribution of AI-related changes remain an open question.
Key Questions
What is the report’s main finding about automation and jobs?
It cites James Bessen’s study of 271 U.S. occupations from 1950 to 2010, in which elevator operator was the only occupation whose disappearance was largely attributed to automation. The report says automation more often changed tasks within jobs.
Did the 1945 elevator workers’ strike cause operators to be replaced?
The strike is documented, but the report says it found no document proving the strike caused building owners to adopt unattended elevators. It describes that connection as a popular story, not an established fact.
How much did the number of typists change?
The source reports 985,000 word processors and typists in 1988 and about 40,400 in 2025. Those figures show a steep long-term decline, but do not alone establish how much any single technology caused it.
Does this history show that AI will eliminate only a few jobs?
No. The historical study covers a different period and earlier technologies. It offers a comparison, not a reliable count or forecast of AI-related job losses.
Source: ThorstenMeyerAI.com
