A deeper look into the impact of new technologies on our work
The case of ATM in the US banking industry
If you’re interested in artificial intelligence, maybe you’ve heard that ‘artificial intelligence is like the electricity of the 21st century’. In other words, many people think that AI has the potential to create such a huge level of change as electricity has since it was discovered about 100 years ago.
In my other article ‘Will artificial intelligence take over our jobs?’, I mentioned it’s more important to understand AI is not going to simply replace existing jobs, but rather augment the performances being carried out by certain employees by automating ‘particular tasks’ in the overall context of business practices. And I presented the case of ATM and (non-) missing bank tellers in the US market as an example.
In this article, I will take a closer look at how the growing adoption of ‘ATM’ impacted the strategic choices in the transformation processes of the US banking industry, and how the nature of the relationship between new technology and the ‘job’ is not a simple ‘replacement’, but more complex function involving cost/value of activities, strategic choices, etc., resulting in the various form including ‘ augmentation’ of tasks, ‘replacement’ of jobs, and ‘emergence’ of new jobs.
New Tech Conference 1998, Denver, Colorado
But before delving into ‘behind-the-scenes’ of US banking industry meeting ATM, let’s turn back time for a second — on March 27th, 1998, in the New Tech 1998 conference in Denver, Colorado.
Here, Neil Postman, a prominent American cultural critic and professor at New York University, gave a keynote lecture. Professor Postman has been a long-time scholar of how new technologies relate to human society, and the book ‘Amusing Ourselves to Death’, a 1985 book that rose to stardom, shows how television technology is destroying public discourse and turning everything into entertainment. I think it has something to do with how we feel about the impact of today’s media and how our lives exposed to it are deteriorating.
Since this book, Professor Postman has strongly criticized the tendency to respond to all social problems through technical solutions.
By the time Postman was on the podium of the New Tech conference in Denver, the world was full of end-century pessimism. Since the 1980s and through 1990s, personal computers have become deeply embedded in our daily lives, and the Internet has begun to spread. By 2000, the so-called Y2K issue was being seriously discussed, with the spreading fear that the huge chaos was waiting for us when it became January 1st, 2000.
So people had ambivalent feelings about the new technology. In the past, Y2K was close to groundless fear, but it was hard to predict what the impact of the widespread use of information technology would be in the future. At the end of the century, we had high expectations for the great technological change ahead but felt that it would not solve the fundamental problems of mankind.
Postman says, “I don’t think the issues that we’re going to face in the 21st century will be any more surprising or complicated than the ones that we’ve been facing for a long time.”
The technologies that are already emerging today and that we will see in the future are already much more complex and difficult to understand than the mechanical technologies of the past, but the old scholar, who has long looked at the relationship between technology and society, has seen the fundamental properties of various technologies.
According to Postman, the new technology is always ambivalent. Just like the deal ‘Faust’ made with the devil ‘Mephistopheles’, the new technology offers a new level of convenience to humans on one hand but takes something in return.
Postman’s point, which seems so obvious to us at first glance, had the advent of computer and Internet technology in 1990 in mind, but we are still repeating similar concerns after 20 years.
In particular, the idea that artificial intelligence is the technology with the potential to replace human intellectual labor, unlike the previous technologies that have contributed to overcoming human physical limitations, lies at the heart of the question, “Will artificial intelligence take over our jobs?”.
The Emergence of ATM and its Impact on US Banking Industry and Bank Tellers’ Job
Clearly, the technology that we call ‘artificial intelligence’ definitely has the potential to not only influence the manual workers but also have an impact on the knowledge workers. Many companies already have started various experiments to use artificial intelligence models in their business processes, new products, and services, and sometimes we see the media bombarding the news on the general public with the threatening title like “jobs lost to AI”.
However, it is worth noting that ‘job’ is something that ‘changes’ over time in a dynamic manner rather than is ‘completely replaced’ instantly by the introduction of new technology.
You all know ATMs that we’re so familiar with today. Younger generations may be more familiar with mobile payment apps than ATMs. ATM was invented by a Scottish inventor named John Shepherd-Barron. John is said to have been in the printing business, and he came up with the idea that there should be a way to withdraw money anywhere, regardless of location, and that he would apply the principles of the vending machines to bank deposits, similar to his favorite then-existing chocolate machines. I don’t know if this story was made by John Shepherd-Barron or after ages — because this story resembles the story of Archimedes’ realization of the principle of buoyancy in the bath and shouting Eureka! — anyway, that’s what they say.
Barclays, the UK-based financial firm, believed John’s idea made sense and hired him. Since then, Barclays invested in the invention of ATM devices and installed the world’s first ATM machine in Enfield, North London, on June 27, 1967.
ATMs have spread rapidly since then through the 1990s and are still widely used. With the prevalence of ATMs, have jobs of bank tellers, or ‘tellers’, disappeared?
The data says they haven’t. Let’s take a look at research about ATM expansion in the US market. In the United States, the first ATM was installed in 1971. Since then, ATMs have increased dramatically, reaching 400,000 units in 2009. At the time, ATMs were strategically used as a means of expanding their businesses to areas where regulation was most difficult so they cannot open new branches easily. But in the process, bankers knew intuitively that ATMs would lower the average costs of branch operation by reducing the number of employees in teller-centric branches. Richard Rosenberg, the vice chairman of Wells Fargo at the time, said the expansion of the ATM-based electronic transaction would not only reduce the number of bank branches, but also dramatically reduce the number of remaining bank employees.
However, in the mid-1990s, the number of tellers has not decreased, but rather increased steadily, amid the rapid spread of ATMs. At this time, the total number of bank branches also increased, but the pattern of the increase was linear. If the full-fledged distribution of ATMs resulted in the elimination of large-scale bank teller jobs, this graph wouldn’t come out.
So why did this result come out? There’s got to be several explanations.
First of all, ATMs actually replaced some of the ‘work’ — or, task — that the tellers were doing. But ironically, it had the effect of creating that much more demand for the bank tellers. By urban area standards, the number of employees needed to run a bank branch decreased from about 20 in 1998 to 14 in 2004, but the cost of opening a new bank branch and expanding its business was also reduced at the same time. The result is a 45% increase in the number of bank branches in the U.S. urban area between 1998 and 2004, offsetting the number of bank tellers reduced by the proliferation of ATMs.
In addition, with the proliferation of ATMs and the increased convenience of banking from the customers’ perspective, the size of banking industry itself increased dramatically. So the need for bank tellers became stronger, the need for bank tellers to deal with ATM’s poor, limited performance and to help customers who preferred doing it face-to-face with bank tellers to using ATM machines.
There are other aspects worth digging into. At this time of ATM proliferation, the regional boundaries of the US banking industry, which had primarily enforced state-by-state regulations in the US market, began to disintegrate. Thus the competition in the banking industry began in earnest as ATM and other technologies demonstrated economies of scale. Naturally, there were fierce battles to secure new customers and protect existing customers among the banks.
Different ‘strategic groups’ emerged in these battles. Some banks saw ATMs as a complete replacement for bank tellers, while others found greater strategic value in increasing revenue per customer by turning the bank tellers into the service and sales teams for customers. We can call the first group’s strategy as ‘cost-based strategy’ and the second group’s strategy as ‘differentiation-based strategy’.
Lessons Learned : The Mechanism between New Technology and Changes in Our Work
There are some implications in the story above for predicting the relationship between artificial intelligence and human occupations.
First of all, as with all other technologies, artificial intelligence is a technology (or a set of technology) with ‘ambivalence’, so is the impact of artificial intelligence on our jobs as Neil Postman pointed out. Whether artificial intelligence will have positive or negative impacts on our society, our professions depends on how all stakeholders guide the development of this technology and how they create a social foundation.
The second implication is that artificial intelligence does not perform a ‘job’ a person can do at a 1:1 exchange value. Replacing a particular ‘task’ with technology necessarily changes the economics of the entire activities in the value chain, thereby increasing the value and demand of the work surrounding the task (complimentary activities). In this process, we will be able to catch a wide-spectrum of opportunities spanning from helping people focus on higher-value activities supported by the automation with artificial intelligence technologies to developing new businesses, services, and products (which will create enormous number of new jobs).
A white paper on ‘artificial intelligence and jobs’ released by Microsoft on March 24, 2020, predicts that it will have a significant impact on new job creation. In particular, it explains that automating repetitive and routine tasks can help workers with their high-dimensional tasks. These changes in workers’ roles are expected to increase workers’ wages by more than 10% by 2030.
Finally, the greater the impact of technology, the more likely this change will take place over a considerable period of time. In the ATM case above, the prediction that ATMs would reduce the bank tellers’ job appeared in the early 1980s, but the prediction is becoming real now in 2020s not just because of the emergence of the technology but because of multiple factors including the changes in regulations, economic growth and culture, and the specific strategic choice made by the banks, the key stakeholders using ATMs.
It is very important to carefully plan and implement the education programs that allow members of society as a whole to adapt to changes coming along with the new technology. In particular, artificial intelligence is so different from the technologies we have adapted over the past few decades that we need a new approach to develop long-term education content, teaching systems, etc.
What we should not forget is that the impact of artificial intelligence does not equitably distribute the benefits and costs to everyone. Every technological change has a winner and a loser. From the industrial revolution that sparked the Ruddite Movement to the 19th and 20th centuries, the issue of unemployment driven by new technology adoption has brought various social problems, in which winners tend to deliberately ignore the ambivalence of technology change and dismiss people who care about short-term side effects as detractors. So those who have been swept away by the so-called “technical unemployment” are, unfortunately, perceived as short-term collateral damage.
The problem is, if unemployment from the emergence of technology in the past has remained at the risk of somewhat limited social groups, the cost of our society to face the fourth industrial revolution, driven by artificial intelligence, may come as a wave that no one can avoid. So, we need to take measures to help workers in the middle and low-skilled professional groups who will be exposed to the difficulties of this change in the short term, as much as to nurture the best researchers, engineers, and entrepreneurship-minded start-ups who will lead the AI era.
We need principles, environment, and strategy to introduce artificial intelligence while protecting their ‘jobs’. For example, the question that companies have to ask themselves in the process of introducing artificial intelligence technology is not “How do we reduce the workforce?” but “How can we bring artificial intelligence to help our people create new values and contribute to our business?. It is also important to actively listen to the voices of workers and create an artificial intelligence system that reflects their needs during this introduction process.
Closing Remarks
The future we’re going to paint with artificial intelligence technology is neither ‘utopia’ nor ‘dystopia’. The etymology of the word ‘utopia’ means ‘nowhere’. If dystopia is a utopian-derived opposite, it probably means something like ‘somewhere or one day maybe’. Maybe that’s why we keep reminding ourselves of the dystopian future that AI technology brings. So we need to put more effort to move forward to the future that is closer to utopia.
Beyond the dystopian confrontation, I hope we can build a future that creates many new social values and new jobs together, through the collaboration between humans and machines.
