🔍 Read the full analysis: Machines At Work: The AI And Human Expertise Paradox on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
ATMs took over routine cash tasks but did not immediately eliminate bank tellers: U.S. teller employment rose modestly from about 500,000 in 1980 to about 550,000 in 2010, while branches expanded. The source argues that mobile banking drove a later decline in teller jobs, but the episode does not establish how AI will affect employment.
ThorstenMeyerAI.com published an account on November 20, 2026, tracing how automated teller machines took over cash transactions without immediately erasing teller jobs—and how mobile banking later coincided with a steep fall in U.S. teller employment. The installment, Episode 18 of “Work After Work,” uses banking to examine a question relevant to current debates about AI: when a machine automates tasks, what happens to the people whose jobs include them?
The episode begins with Barclays’ first cash machine, demonstrated by actor Reg Varney outside a branch in Enfield, north London, on June 27, 1967. According to the account, the early machines were built with banknote printer De La Rue and used a special paper cheque marked with carbon 14, along with a personal code. Plastic cards were not yet in use. The machine could dispense cash after bank branches had closed, when customers previously needed to visit during office hours to obtain it.
The account says the technology spread internationally: a cash machine opened in Tübingen, Germany, on May 27, 1968, and Chemical Bank installed a Docutel dispenser in Rockville Centre, New York, in September 1969. In the United States, ATM numbers grew from about 100,000 in 1995 to about 400,000 in 2010, according to Bank for International Settlements figures used by economist James Bessen. Yet U.S. teller employment rose modestly, from about 500,000 in 1980 to about 550,000 in 2010, based on Bessen’s count as summarized by MIT economist David Autor.
The episode attributes that outcome to a change in the economics and organization of banking. Fewer tellers were needed at a typical urban branch—about 13 rather than 20 between 1988 and 2004—but lower operating costs helped banks open more branches. The number of branches in U.S. cities rose 43% from 1988 to 2004. Later, customer use shifted toward phones: in 2023, 48.3% of U.S. households with a bank account mainly used mobile banking, while 15.1% mainly used a teller. Bureau of Labor Statistics figures cited in the episode put U.S. teller jobs at about 600,500 in 2008 and 339,200 in 2025.
Infographic 01 · The paradox
Four times the machines, more tellers
ATMs took over the teller's core task. For three decades, banks answered with more branches, not fewer tellers.
US commercial bank offices (head offices + branches) · year-end · FDIC
Teller and ATM figures: Bessen’s analysis of Census and American Community Survey microdata, as summarised by Autor (2015); ATM counts from the Bank for International Settlements. Branch growth in cities: Bessen (2015), urban markets. FDIC series: FDIC-insured commercial banks in the US states and DC; savings institutions are not included.
Sources: D. Autor, Journal of Economic Perspectives 29(3), 2015; J. Bessen, Toil and Technology, IMF Finance & Development (March 2015); FDIC, Historical Bank Data (commercial banks, year-end).
Powered by Thorsten Meyer AI
How Automation Reshaped Teller Jobs
The history complicates a simple forecast that more machines automatically mean fewer workers. ATMs handled important teller tasks, especially dispensing cash and taking deposits, but the episode’s figures show that teller employment did not fall during the period of rapid ATM growth. Lower staffing needs per branch were offset by a larger number of branches, while tellers’ work shifted toward customer interaction and selling financial products.
The later decline matters because it points to a different mechanism: customers moved away from branches as mobile banking spread. That distinction is relevant to AI discussions. Automating a task may change a job without eliminating it, but whether employment holds or falls can depend on demand for the service, how organizations use cost savings, and whether customers adopt a different channel. The banking example is historical evidence, not a direct forecast of AI’s effects across other industries.
Products Worth Considering

Samsung 27-Inch Odyssey G5 QHD Gaming Monitor (G50F)

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

Jackery Explorer 2000 v2 Portable Power Station, 2042Wh, 2200W
As an affiliate, we earn on qualifying purchases.
From Cash Machines to Mobile Banking
A bank teller’s traditional duties included handing out cash, accepting deposits, cashing cheques and answering customer questions. The ATM was named for the occupation and automated part of that work. The episode draws on Bessen’s analysis of teller employment and branch growth, and on Autor’s 2015 summary of that research. It describes the changing role as a move from cash handling toward relationship banking, in which staff advise customers and offer products such as loans and credit cards.
The episode also reports a later contraction in physical banking. In Germany, bank branches fell from 45,467 at the end of 2004 to 16,799 at the end of 2025. It says cash machines are now retreating in Germany, China and Japan, although it provides no comparable counts for those ATM declines. These figures place the U.S. employment change within a broader shift toward digital banking, while not proving that technology alone caused every reduction.
“actually rose modestly from 500,000 to approximately 550,000 over the 30-year period from 1980 to 2010”
— David Autor, MIT economist, summarizing Bessen’s findings in 2015
What the Banking Figures Cannot Show
The episode presents employment and banking-use figures across different periods and measures. It does not establish that mobile banking alone caused the decline in teller jobs between 2008 and 2025, or separate the effects of branch closures, changing customer preferences, staffing decisions and other industry changes. The historical relationship between ATM growth and teller employment also cannot, by itself, predict what AI will do to jobs in banking or elsewhere.
The source gives no detailed counts or time series for the reported retreat of cash machines in Germany, China and Japan. It also does not provide a forecast of how many teller roles may remain, or how quickly banks will adopt AI for customer service and other tasks. Those points remain open.
The Next Test for Banking Work
The episode’s next analytical step is to compare the ATM era with the shift to mobile banking and then consider what those changes can—and cannot—say about AI. The published figures show a historical sequence, not a conclusion about future employment. Any assessment of AI’s effects will need current evidence on which banking tasks are automated, how customer demand changes, and whether banks reinvest savings in new services or reduce staffing.
For readers tracking the labor impact, the clearest next indicators are future teller employment data, branch counts and measures of how customers use digital and in-person services. Until those are available, the account supports a narrower takeaway: automation changed the teller’s job first; later digital adoption accompanied a substantial reduction in the number of teller positions.
Key Questions
Did ATMs immediately eliminate bank teller jobs?
No. The figures cited in the episode show U.S. teller employment rising modestly, from about 500,000 in 1980 to about 550,000 in 2010, even as ATM numbers grew rapidly. The account says the number of tellers needed per branch fell while the number of branches increased.
What happened to teller work as ATMs spread?
ATMs took over routine cash tasks, while the remaining role shifted toward customer assistance and product advice, described in the episode as relationship banking. The extent of that change likely varied by bank and period; the source does not give a complete job-task breakdown.
When did U.S. teller employment begin to fall?
The source gives figures showing a decline from about 600,500 teller jobs in 2008 to 339,200 in 2025, using Bureau of Labor Statistics counts. It links the later period to the rise of mobile banking but does not establish a single cause for the decline.
Does this history prove AI will create or eliminate jobs?
No. The episode uses banking history to show that automation can change tasks and job design, but it does not provide evidence that settles AI’s employment effects. Outcomes may depend on customer demand, adoption and how employers use the savings from automation.
Source: ThorstenMeyerAI.com
