The Unexpected Tension Between AI And Human Expertise
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: The Unexpected Tension Between AI And Human Expertise on ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on monitors, keyboards and dev gear

  • Business-only prices and quantity discounts
  • Tax-exempt purchasing
  • Multiple users, one account, clear invoices
As an affiliate, we earn on qualifying purchases.

TL;DR

The account traces how ATMs took over cash-handling tasks without initially eliminating bank teller jobs: US teller employment rose modestly from about 500,000 in 1980 to about 550,000 in 2010, as cheaper branches and customer demand supported growth. Later, mobile banking and branch closures coincided with a steep fall in US teller employment, illustrating how automation can reshape work before reducing demand for workers.

A November 20 analysis of the automated teller machine’s effect on bank work argues that the technology first changed what tellers did rather than eliminating their jobs, while later growth in mobile banking coincided with a sharp drop in teller employment. The history offers a concrete example of why automating a task does not always remove a job immediately—and why the longer-term effect can change as customers adopt new ways to bank.

The article recounts the first Barclays cash machine demonstration outside a branch in Enfield, north London, on June 27, 1967. Actor Reg Varney withdrew cash from a machine that could dispense up to ten pounds in one-pound notes. Early machines did not use today’s plastic cards: Barclays’ first machines accepted a special paper cheque marked with carbon-14 and required customers to enter a personal code. Before such machines, customers generally had to visit a bank during business hours to get cash, according to the bank archive cited in the account.

In the United States, the number of ATMs rose from about 100,000 in 1995 to about 400,000 in 2010, according to figures attributed to the Bank for International Settlements and economist James Bessen. Yet US teller employment increased modestly over a longer period: from about 500,000 in 1980 to about 550,000 in 2010, as MIT economist David Autor summarized Bessen’s findings. The figures cover different periods and should not be read as a direct year-by-year comparison.

Bessen’s explanation is that ATMs reduced the number of tellers needed at each branch, but made branches cheaper to operate. In US cities, the branch count rose 43% between 1988 and 2004, while a typical branch’s teller staffing fell from about 20 to about 13. The article says tellers shifted away from routine cash handling toward customer service and product advice, often described by banks as relationship banking.

At a glance
analysisWhen: Published November 20, 2026; historical…
The developmentA November 20, 2026, historical analysis uses the ATM and subsequent mobile banking to examine how automation changed bank teller work and eventually reduced teller employment.
Thorsten Meyer AI
Work After Work · Ep. 18

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.

×4ATMs in the US, about 100,000 (1995) to 400,000 (2010)
≈ +10%US bank tellers, about 500,000 (1980) to 550,000 (2010)
20 → 13tellers per branch in a typical US city, 1988 to 2004
+43%bank branches in US cities, 1988 to 2004

US commercial bank offices (head offices + branches) · year-end · FDIC

  • 198052,879
  • 198858,937
  • 1995ATMs start to spread fast65,578
  • 200477,302
  • 2012peak88,522
  • 202572,658

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 account matters because it separates the automation of individual tasks from the disappearance of an occupation. ATMs could dispense cash and accept deposits, but branches still needed employees for customer questions and other services. Lower operating costs and additional branches helped offset the reduced staffing required at each location.

The later figures point to a different phase. In 2023, 48.3% of US households with a bank account mainly used mobile banking, while 15.1% mainly used a teller, according to the source. US teller employment fell from about 600,500 in 2008 to 339,200 in 2025, based on Bureau of Labor Statistics figures cited in the account. The figures show a decline over that period, but do not by themselves establish how much mobile banking, branch closures or other factors caused it.

For readers weighing claims about AI and work, the example is a useful caution against simple forecasts. New technology can change the mix of tasks, the cost of providing a service and the amount of customer demand. Those effects may pull employment in different directions, and the outcome can shift as technology and customer habits change.

Ninja Foodi Smart 6-in-1 Indoor Grill and Air Fryer, 4 qt, FG551

Ninja Foodi Smart 6-in-1 Indoor Grill and Air Fryer, 4 qt, FG551

Breville Barista Express Impress Espresso Machine with Grinder, Brushed Stainless Steel

Breville Barista Express Impress Espresso Machine with Grinder, Brushed Stainless Steel

As an affiliate, we earn on qualifying purchases.

From Cash Machines to Mobile Banking

The historical account places the Enfield demonstration at the start of a broader rollout. Germany’s first cash machine opened in Tübingen on May 27, 1968, according to the source, and Chemical Bank installed a Docutel cash dispenser in Rockville Centre, New York, in September 1969, as dated by the Smithsonian and Computer History Museum. The machines took on a visible part of teller work, but did not initially replace the full range of services provided at a branch.

US teller employment and branches later moved in different directions. The source reports that the number of US commercial bank offices grew from 58,937 at the end of 1988 to 77,302 at the end of 2004, according to the FDIC count it cites, and reached 88,522 in 2012. It also attributes branch growth in that period partly to deregulation that allowed banks to open branches more freely.

The account describes a subsequent shift to phone-based banking. Germany’s branch count fell from 45,467 at the end of 2004 to 16,799 at the end of 2025. It says cash machines are also retreating in Germany, China and Japan, though it does not provide counts or a detailed explanation for those changes.

“A bank teller who can tally currency but cannot provide ‘relationship banking’ is unlikely to fare well at a modern bank.”

— David Autor, MIT economist, summarizing Bessen’s findings in 2015

What Drove the Later Decline

The cited employment and banking-use figures show that teller jobs declined while mobile banking became common, but the material does not quantify the separate effects of mobile apps, branch closures, demographic change or other industry shifts. It also does not establish whether the teller-employment decline will continue at the same pace.

The article’s historical account is an analogy for current debates about AI and human expertise, not evidence that AI will produce the same labor-market pattern. The source does not provide a current study of AI adoption, measured employment effects or specific occupations affected by AI. Those questions remain open and would require evidence beyond the ATM example.

The Next Test for Banking Work

The next development to watch is whether banks continue to reduce physical branches and teller staffing as customers rely more on digital services, or whether new services create work requiring human support. The source gives historical employment and usage figures but does not set out a forecast or announce a specific next milestone.

For the wider AI debate, the relevant test is not only whether a system can perform a task, but whether its use changes demand for the service, the number of workplaces and the work left for people. Future employment data and task-level studies will be needed to determine how that balance plays out in sectors adopting AI.

Key Questions

Did ATMs immediately eliminate bank teller jobs?

No. The account says US teller employment rose modestly from about 500,000 in 1980 to about 550,000 in 2010, even as ATM numbers roughly quadrupled from 1995 to 2010.

Why did teller employment rise while ATMs spread?

According to economist James Bessen’s explanation in the source, ATMs reduced staffing needs at individual branches but lowered branch operating costs. The number of branches in US cities rose 43% from 1988 to 2004, offsetting some of the staffing reduction per branch.

When did teller employment begin to fall sharply?

The figures cited in the account show US teller employment falling from about 600,500 in 2008 to 339,200 in 2025. The account links the later period to mobile banking and branch contraction but does not isolate each cause.

Does the ATM story predict what AI will do to jobs?

Not by itself. It shows that automating tasks can reshape an occupation before employment falls, but the source provides no direct measurement of AI’s labor-market effects. AI outcomes will need to be assessed with evidence from the affected industries and workers.

Source: ThorstenMeyerAI.com

You May Also Like

Why is Doordash not working? DoorDash down for many Sunday

Many users are experiencing issues with DoorDash on Sunday, with the platform reportedly down for a significant number of users. Details are still emerging.

Is There Evidence Jensen Huang Plans To Leave?

The cited reports say Jensen Huang resigned from Nvidia on Nov. 11. They do not establish a coordinated political effort or prove espionage allegations.

Threlmark: Disk Is the Contract

Threlmark launches a new approach where the roadmap is a plain JSON file on disk, enabling open, interoperable, and durable project planning.

Nvidia And The Spy Probe: What Reports Say About Trump And The Board

Nvidia confirmed a security investigation, while media reports say its board discussed Jensen Huang’s future. No decision has been announced.