🔍 Read the full analysis: From ATMs To AI: Why Automation Has A Human Twist 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
The history of ATMs shows that automating a task did not immediately erase the bank teller job: U.S. branch growth helped offset the need for fewer tellers at each branch. Later, mobile banking and branch closures accompanied a steep fall in teller employment. The account frames AI as another automation wave, but does not provide evidence that current AI systems will produce the same employment pattern.
A November 20 analysis of ATM and bank teller history argues that the first wave of cash machines reduced the number of tellers needed at each U.S. branch but did not initially shrink the occupation overall. The account’s central point is that automation changed the work, while later mobile banking and branch closures coincided with a much larger decline in teller employment.
The article traces the first Barclays cash machine to June 27, 1967, when actor Reg Varney withdrew up to ten pounds in one-pound notes from a machine outside a branch in Enfield, north London. Because customers did not yet have plastic cards, the early Barclays machines accepted a special paper cheque marked with carbon-14 and required a personal code, according to the account. A German cash machine opened in Tübingen in May 1968, followed by a Docutel dispenser at a Chemical Bank branch in New York in 1969.
In the United States, the number of ATMs rose from about 100,000 in 1995 to 400,000 in 2010, a roughly fourfold increase, using figures attributed to the Bank for International Settlements and economist James Bessen. Yet teller employment increased modestly over the longer period: MIT economist David Autor’s 2015 summary of Bessen’s findings put the count at about 500,000 in 1980 and 550,000 in 2010. Those figures do not mean teller work was unaffected. In U.S. urban markets, the typical number of tellers needed per branch fell from about 20 to about 13 between 1988 and 2004.
The article says lower branch operating costs helped banks expand their networks, offsetting some of the staffing reduction at individual locations. It reports a 43% increase in U.S. city branches from 1988 to 2004. Later, the balance changed: Bureau of Labor Statistics figures cited in the source show U.S. teller jobs falling from about 600,500 in 2008 to 339,200 in 2025. In 2023, 48.3% of U.S. households with a bank account mainly used mobile banking, compared with 15.1% that mainly used a teller.
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 Work
The account offers a useful distinction for readers weighing predictions about AI and employment: a technology can automate specific tasks without immediately eliminating the job built around them. ATMs handled routine cash transactions, while tellers increasingly focused on customer interaction and sales. Bessen described the shift as cash handling becoming less important and human interaction more important.
That history does not establish what AI will do to today’s jobs. It does show why counting automated tasks or machines alone can give an incomplete picture. Employment can also depend on demand for the service, the cost of providing it, the number of locations, and the tasks workers take on after automation. In banking, branch expansion initially helped sustain teller numbers; the later move to mobile banking and branch closures accompanied a substantial employment decline.
For customers, the change also affects how they access help. Digital channels can make routine transactions available without a branch visit, while reducing the number of staffed locations and face-to-face interactions. The supplied figures document these shifts but do not quantify their effects on service quality, access, or individual workers’ transitions into other roles.
Popular right now

Shark PowerPro Reveal Plus Cordless Stick Vacuum, Charcoal (IZ376H)

Sony WH-1000XM6 Wireless Noise-Cancelling Headphones – Midnight Blue

KitchenAid 5.5-Quart Bowl-Lift Stand Mixer, Porcelain White
As an affiliate, we earn on qualifying purchases.
From Cash Machines to Mobile Banking
The first ATM addressed a specific constraint: before cash machines spread, getting cash generally meant visiting a bank during office hours. Early devices did not match the ease of modern card-based withdrawals; Barclays’ first model relied on a marked cheque and a code. The machine’s name—automated teller machine—reflected its connection to a human role, even though it performed only some teller tasks.
The U.S. employment comparison covers different periods and measures, which should not be conflated. The 1980–2010 headline count summarized by Autor is not the same as Bessen’s full-time-equivalent estimate. The latter, based on Census and American Community Survey microdata, counts hours worked as equivalent full-time positions; the source says that measure rose by 2.0% annually from 2000 to 2013. The article also cites FDIC data showing U.S. commercial bank offices rising from 58,937 in 1988 to 77,302 in 2004, then peaking at 88,522 in 2012.
International examples point to continued changes in physical banking. The source reports that Germany’s bank branches fell from 45,467 at the end of 2004 to 16,799 at the end of 2025, and says cash machines are also retreating in Germany, China and Japan. It does not provide comparable ATM counts for those countries.
““A bank teller who can tally currency but cannot provide ‘relationship banking’ is unlikely to fare well at a modern bank.””
— David Autor, summarizing Bessen’s findings in 2015
What the Historical Figures Cannot Show
The supplied account presents historical figures and an interpretation of how branch growth offset lower staffing per location. It does not establish that ATMs alone caused branch expansion or that mobile banking alone caused the later job decline. Other factors, including banking regulation and wider changes in how customers use financial services, also matter; the source specifically notes deregulation as one factor in branch growth.
The figures also use different time spans and definitions. The 1980 and 2010 teller counts, the full-time-equivalent estimate for 2000–2013, and the BLS job counts for 2008 and 2025 should be read as separate measures rather than a single continuous series. The account supplies no detailed breakdown of who left teller work, what roles they moved into, or how regional access to branches changed.
Although the article’s title connects automation’s history to AI, its evidence concerns banking machines, branch networks and mobile services. It does not present data on current AI adoption, AI-related job losses, or a forecast for particular occupations. Whether AI will raise demand for human work in the way ATM-enabled branch expansion once did remains unresolved by this material.
The Next Test for AI and Work
The historical account closes with a question that applies to current automation debates: which tasks will change, and what work will remain? Answering it for AI will require evidence beyond the number of systems adopted. Relevant indicators include employment over time, changes in job duties, hiring and training, wages, and whether lower service costs expand demand enough to create or sustain other roles.
For banking, the next developments to watch include whether branch and cash-machine networks continue to contract, how customers divide transactions between digital and in-person services, and whether banks retain human staff for advice and complex requests. The source provides figures through 2025 for U.S. teller jobs and German branches, but no forecast or subsequent employment data. Its account therefore describes a historical pattern, not a settled outcome for AI.
Key Questions
Did ATMs eliminate bank teller jobs?
Not immediately in the United States, according to the figures cited in the article. Teller employment rose modestly from about 500,000 in 1980 to 550,000 in 2010, even as the number of ATMs roughly quadrupled from 1995 to 2010. Tellers per branch fell, while branch numbers grew.
Why did teller employment initially hold up?
The account says ATMs made branches cheaper to operate, helping banks open more locations. That expansion partly offset the reduction in tellers needed at each branch. It also points to deregulation that allowed banks to open branches more freely.
When did U.S. teller jobs start falling sharply?
The source gives a decline from about 600,500 jobs in 2008 to 339,200 in 2025, based on Bureau of Labor Statistics figures. It links the later period with mobile banking and branch closures, but does not isolate the effect of any one cause.
Does ATM history predict what AI will do to jobs?
No. It illustrates how technology can change tasks and job duties, but the supplied evidence does not measure AI’s effect on employment. Outcomes may depend on demand, costs, organizational choices and the work people take on after automation.
Source: ThorstenMeyerAI.com
