Eyes glaze over when I mention my interest in researching automated teller machines. Yet after I explain why I think they're relevant, many people can easily recall personal anecdotes in which an ATM plays a central role: a chance encounter with a long-lost friend while waiting in a queue, or the fear of being robbed in an unfamiliar location, or the feeling of seeing an insufficient funds notice displayed on the screen.

Most urbanites have interacted with the ubiquitous "cashpoint." Paul Volcker, of the U.S. Federal Reserve fame, even considered it the "only useful innovation in banking." Cashpoints appear frequently on TV and in printed news because, for most consumers, they're one of the few points of contact with today's otherwise-ephemeral financial services.

In spite of their cultural significance, ATMs recede into the noise of everyday memory. Few stop to reflect on how they—and the computer infrastructure that supports them—became the backbone of contemporary retail payments.

The cash dispenser was born almost 50 years ago, in 1967. For many, this was the first tangible evidence that retail banking was changing; the introduction of the ATM marked the dawn of contemporary digital banking. Several lay claim to the invention of the cashpoint, including John Shepherd-Barron and James Goodfellow in the U.K.; Don Wetzel and Luther Simjian in the U.S.; and even engineering companies like De La Rue, Speytec-Burroughs, Asea-Metior, and Omron Tateisi. But the ATM is a complex technology. There was no single eureka moment that marked its arrival.

The ATM finds its origins in the 1950s and 1960s, when self-service gas stations, supermarkets, automated public-transportation ticketing, and candy dispensers were popularized. The first cash machine seems to have been deployed in Japan in the mid-1960s, according to a Pacific Stars and Stripes account at the time, but little has been published about it since. The most successful early deployments took place in Europe, where bankers responded to increasing unionization and rising labor costs by soliciting engineers to develop a solution for after-hours cash distribution. This resulted in three independent efforts, each of which entered use in 1967: the Bankomat in Sweden, and the Barclaycash and Chubb MD2 in the U.K.

Cashpoints materialized thanks to a long chain of innovations. Some were of a general nature, such as steel, video-display units, plastic, magnetic tape, or (more recently) the Windows operating system. Others were purpose-made, such as the cash output mechanism and, in the 1960s, the previously non-existent algorithm that associated an encrypted PIN with a customer account. These components were developed through active collaboration between groups of bankers and engineers, each of which attempted to solve different aspects of the complex challenges inherent in the development of the ATM.

Never before had electronic equipment been so exposed to the elements. The necessity of human intervention in early systems invited further automation. For instance, they could easily jam or run out of product. They could erroneously dispense several bank notes instead of just one—all without the owner's knowledge. They were activated by plastic or paper tokens that would only activate for the operating bank and, in some cases, only that particular bank location. Some banks would keep the token in the machine and return it to the customer (by post) once the account had been debited. As a result, early ATMs were standalone, clunky, unfriendly, and inflexible. They could do one thing: dispense cash when activated by a token.

Given these constraints, it's not surprising that it took more than a decade for banks to deploy cashpoints beyond a handful of experiments. In its early days, few believed that the cashpoint would make a difference to the average consumer. In context, this prediction might have seemed sure; cashpoints appeared before credit or debit cards were a popular alternative to bills and coins, at a moment in time when most of the world's citizens worked in a cash economy. With the exception of the U.S. and France, even personal checks were largely limited to the wealthy.

Updating central records from the point of a transaction is easy in today's world of mobile banking and e-commerce, but the cashpoint was one of the first devices to use real-time networking. Early in the ATM's development, creating a way to communicate with a central computer (and therefore inform customers of their account balances) became an overriding design concern. In cooperation with IBM, Swedish savings banks began testing a networked cashpoint in 1968. A collaboration between IBM and Lloyd's Bank followed, and that bank deployed several networked devices in the United Kingdom in 1973. But widespread online authorization still had a long way to go. Throughout the 1970s, IBM engineers developed the rails, pipes, and standards on which other elements of the payments ecosystem (such as credit cards and point-of-sale terminals) would eventually depend.

The ATM freed the average consumer from lengthy queues for services that had previously been limited to bank hours. As devices spread, this convenience steadily changed patterns of consumption, enabling unplanned weekend shopping and impromptu dining. At the same time, it allowed retail banks to grow their customer base by granting access to consumers who'd previously been excluded from using a current account or a credit card. The nature of work in bank branches also changed as employees relocated away from teller services and into sales. High-margin services and products like car insurance, credit cards, investment funds, and mortgages owe a debt to the outsourcing of ordinary banking to ATMs. But when such sales opportunities failed to materialize, banks also began to cut costs by reducing branch staff and closing down branches. This process continues even today, with so-called "branch transformation" remaining a hot topic in the industry.

Bank regulators across the world have actively shaped ATM technology by dictating who can own and operate them, monitoring the cost of withdrawals as well as where they can be physically located. But the average person has also influenced ATMs: the way they look, the way they work, and their role as a platform for today's plethora of balance inquiries, deposits, transfers, and (in some European countries) airtime top ups for pay-by-minute cellphones.

Arnd Wiegmann/Reuters

In 1971, a handful of years after the first machines appeared in England and Sweden, manufacturers were operating in Britain (Speytec-Burroughs), the U.S. (Docutel and Diebold), and Japan (Omrom Tateisi). Together, they deployed cash machines in their home countries and across Europe, Canada, Israel, Cyprus, and Latin America. However, by the early 1980s, pioneers such as Chubb, De La Rue, Docutel, and Asea-Metior had left the industry as each failed to keep up with developments in computing and electronics. Other manufacturers, such as Burroughs, hadn't achieved their deployment targets. Citibank abandoned plans to commercialize its proprietary CAT-1 and CAT-2 devices and, instead, continued to use them in its global, proprietary network until the 1990s.

Not so with IBM, which had the marketing muscle, engineering expertise, and business contacts to dominate the market. The company seemed poised to overwhelm its competitors until executives decided to deploy a new model—the IBM 4732 family—which were incompatible with previous models, including the already-successful and widely deployed IBM 3624. Many banks evaluated the machine and refused to buy it because, in a stroke, IBM had made the banks' significant capital investments in the older computer infrastructure obsolete. This obsolescence extended beyond the physical devices inside bank branches to the machines and software that supported communication across the bank's network, and even to standards for shared cashpoint networks. IBM's move soured banks, inadvertently, opening the ATM market to new cashpoint manufacturers. Eventually, IBM abandoned payment-technology systems entirely.

Around this time, two Ohio-based companies, NCR and Diebold, were working on technology that would enable them to dominate the supply of cashpoints for the next two decades. As a result of the IBM 4732 fiasco, NCR built its business on software that emulated the IBM 3624. Meanwhile, IBM and Diebold formed a joint venture in 1984, called InterBold. Its aim was to unite Diebold's self-service technology with IBM's global distribution system. Seven years later, and in spite of growing sales, the joint venture ended: Diebold hadn't achieved the international market breakthrough it'd hoped for and IBM's returns fell short of its expectations, in part due to the growth in local processing architectures, which had invalidated IBM's strategy to link ATMs to its expensive mainframes.

NCR and Diebold were instrumental in turning the cash-dispenser dinosaur into today's sleek, multi-function ATM. The companies' innovations included customer-friendly video display units, programmable buttons alongside the screen, a shift toward dispensing cash horizontally (which reduced jams), and expanded functionality, including money transfers and balance inquiries.

But NCR and Diebold were not alone. Growth in the number of banks deploying ATMs across the world promoted an increase in the number of manufacturers: Honeywell in the U.S.; Phillips, Olivetti, and Siemens-Nixdorf (today, Wincor) in Europe; and Fujitsu, GRG, Hyosung, and Hitachi in Asia. Large European banks also developed proprietary networks, numbering in the thousands of ATMs, which U.S. banks favored shared networks (and their subsequent interconnection fees).

Despite innovations in modular manufacturing, speedier ways to identify delinquent accounts, and the associated reduction in service costs, however, ATMs remained a significant capital investment. The use of dedicated telephone lines limited them to bank branches or high volume non-bank locations, such as busy train stations and big airports. This limitation finally lifted with the advent of digital telephony and the industry's adoption of the Windows operating system. These two seemingly simple modifications transformed the ATM, enabling remote diagnostics and integration with credit card clearance networks. They also enabled the advent of the Independent ATM Deployer (IAD)—ATM vendors unaffiliated with a major financial institution —and renewed growth in the machine's deployment in the late 1990s.

Still, not everything is rosy for the ATM industry. In a cost-reduction move in 2014, for instance, Chilean banks reduced the size of their ATM fleets (as well as the frequency of cash resupplies for existing machines) while encouraging the use of government-sponsored cash remittance networks in mom-and-pop retail stores. This move led to public outcry and anti-bank campaigns on social media. The success of mobile banking in Africa has also created doubt about the need to deploy ATMs in developing countries. Mobile banking and remittances—which alleviate the need for cash and bank branches in rural areas—offer the chance to increase financial inclusion in Africa, Asia, and Latin America while obviating the substantial investment required to install and maintain proprietary ATM networks. Despite these advantages, the fate of mobile banking and remittances, for many developing countries, remains uncertain.

From its humble and uncertain beginning nearly 50 years ago, the ATM has become pervasive. But it wasn't until the 1980s, more than 15 years after the machine's invention, that the ATM's success was assured. Today, we're asked for our PINs in libraries, on the Internet, and in every sort of retail store, for which debit cards have become the de facto currency. The near-total global integration of ATM networks means that we can travel almost anywhere in the world with just a piece of plastic in our pocket, confident that we'll have access to cash in places as far afield as Hong Kong, Easter Island, Giza, Paris, and even Antarctica. Some machines now act as Internet kiosks, while others display an advertising by third parties or allow users to purchase minutes for their mobile phones. Yet for all its digital innovations, the quick dispensing of physical cash remains the core, transformative function of the ATM.

An ongoing series about the hidden lives of ordinary things