From a box that gave out cash to a banking terminal
The earliest cash machines of the late 1960s were, in a sense, clever dispensers. What turned them into the ATMs people use today was a set of ideas layered on top: a way to prove who you are, a way to check an account, and a network connecting many machines to many banks. This article follows that second story. The question of which machine opened first, and who gets credit, is treated separately; here the focus is on what the technology did and how it spread.
Competing claims, briefly
Several claims overlap. A machine installed at a London branch in June 1967 is the one most commonly called the first. Swedish and American machines followed within months or a couple of years, and a Scottish engineer filed a patent on a card-and-PIN approach in the mid-1960s. These claims are often presented as contradictory, but they usually describe different features. One was first to open, another first to use a PIN with a plastic card, and a third to be designed as a standing public terminal. The sensible reading is that the ATM emerged from parallel work, and that its core features were assembled over several years.
Identification: the card and the PIN
The first British machines used vouchers that the machine read, along with a number keyed in by the customer. This made for a kind of two-factor check: something you hold, and something you know. Plastic cards replaced vouchers quickly, because a card could be reused and was easier to manufacture at scale.
A card needs a way to store data. Early cards used punched holes or printed patterns, and then magnetic stripes. The magnetic stripe, developed in the 1960s with a team at IBM among those usually credited, put a strip of magnetic material on a plastic card. The card carried an account number, and the machine read it as the card was inserted or swiped. Because a stripe can be copied by someone with the right equipment, the PIN was an important second barrier.
The PIN was never stored in a simple form. In many systems, the PIN was derived or checked using a cryptographic calculation involving the account number and a secret key, so that no one handling the card or the machine could simply read the PIN. Designing this was a notable early use of cryptography in consumer finance, and it prefigured the cybersecurity concerns that later applied to all online services.
Offline machines and online machines
Early ATMs often worked offline. They had a limited ability to verify that a card was valid, and they sometimes kept a list of stolen or lost cards, a local record. A customer could often withdraw only a set amount, and the bank recorded the transaction afterward. This meant a bank was exposed to the risk that a customer would withdraw more than their balance.
The next stage was the online machine, linked to a central computer. In the 1970s, banks began to connect machines to their data centers over telephone lines, using systems that reflected the growth of early computer networks. The central computer could check the balance, record the withdrawal immediately and decline a request that exceeded the account’s limit. Some historians point to IBM-built machines used by British banks around the early 1970s as an example of this transition, though exact dates differ among sources.
Online machines could do more. They could show a balance, accept a deposit or offer a transfer between accounts, and so the ATM stopped being a single-purpose cash dispenser and became a general banking terminal.
Spreading through banks
The spread of the ATM was driven by a mix of economics and customer behavior.
- Cost and staffing. Banks that wanted to cut the cost of handling routine cash withdrawals at counters found an attractive alternative in machines that could work around the clock.
- Competition. Once one bank in a town installed machines, others felt pressure to follow, as customers valued the convenience.
- Location. Machines moved from the walls of branches to shopping centers, airports and gas stations, which broadened their use.
- Shared networks. At first, a customer could use only the machines of their own bank. In the 1970s and 1980s, banks formed shared networks so that one card could work at another institution’s machine. Regional and international networks followed, which made it possible to withdraw cash in a foreign country from a home account.
Shared networks were not only a technical achievement but an agreement on standards: message formats, fees and how banks would settle the amounts owed among them. Those agreements account for much of why a card issued in one country can usually be used in another today, even though the details are not always smooth.
Effects on banking
The ATM changed habits. Customers who once had to withdraw cash during opening hours could do so at nearly any time. Some banks reported that the machines shifted more routine work out of branches, although the effect on the number of branch employees over time is debated, with some studies suggesting branches took on more advisory work. The truth likely varies by country and period, so any single statement should be treated with caution.
Fraud also changed. Thieves devised ways to steal cards or PINs, including card-skimming devices attached to machines, and banks responded with new card technology, such as chips, and improved monitoring. The history of the ATM includes a continuing contest between convenience and security.
A durable terminal
Later developments added chip cards, contactless features and cash recycling, in which a machine accepts deposits and reuses them. Many people now use cards and phones for payments more than cash, which raises the question of the future of the ATM. In the 2020s, the machines remain widespread, but the pattern of use differs by country.
The lasting lesson of the ATM story is that the visible box is the smaller part. Its significance came from the identification system, the link to a bank’s records and the shared networks that made one card work in many places.