An Engineering Conundrum
What will we do when machines do everything?
By Aditya Johri
On a recent research trip to India I fell into conversation with the owner of a small shop whose primary business was serving mobile phone customers. In a country where most customers pay as they go for their mobile phone use, he helps customers pick from among dozens of available plans while also selling handsets. I was in India collecting data for a study on how mobile technology is allowing people without bank accounts to become a part of the formal financial system. Access to banks is something we often take for granted in the West, but a majority of the world’s population is still unbanked. This means people are deprived of regular savings, credit at competitive rates, and an easy mechanism to transfer money. It costs more for the unbanked—who are poorer to begin with—to obtain credit and send money to others.
Until lately, the shop owner was an important intermediary in the same mobile ecosystem that affords his unbanked customers a modest boost in their financial clout. But when I asked how technology was changing his business, he said digitization, especially online transactions, had increased the use of mobile phones overall but had actually harmed his livelihood. Customers could now directly recharge their phones online, without using him as the middle man. The more comfortable they became with online transactions, the less they came to his shop. Increasingly, he said, digitization and online transactions were hurting small-business owners, cutting revenues for many of them by up to 50 percent. “If I don’t do this,” he said, “what will I do?”
His question captures the essential dilemma of a world that is digitizing at such an exponential pace that even people, like the shop owner, with a niche in the technology industry see a threat to their jobs. What will happen when a lot of what people do now is done by machines (in some form or another)? Many scholars and strategists see this problem as one of technological development taking its natural course and argue that technology will also provide the solution. They point to the mechanization of the farm and the industrial revolution as similar phases of technological development that eventually generated more jobs. Others predict fundamental change as digitization and computation software allow machines to perform actions and to respond to inputs in ways that were inconceivable not long ago. This means less dangerous work will be performed by humans, but humans will have a lot less to do. Even if new challenges come up, machines will learn to deal with them.
We are in the initial phase of this transformation. Right now, creation of jobs through services such as Uber and Lyft and of revenue through Airbnb look like positive developments. Yet, in the near future, when self-driving cars and trucks become commonplace, a lot of these jobs will cease to exist. Retail and warehouse sectors will need a lot fewer people than are hired now.
The solutions offered to date are mostly along the lines of training—especially in technical skills, so that people can take advantage of emerging opportunities and prepare for ones that we cannot foresee at the moment. The other solution, the subject of an experiment in Finland, is to provide everyone a basic income so that basic needs are fulfilled and a lack of employability is not a burden. These are acceptable solutions to some degree, but they do not take into account that billions of people across the world, especially in Asia and Africa, are really young—for instance, 50 percent of India’s population is less than 25 years old—and are looking not just for jobs but a purpose in life.
The paucity of thinking around this topic is a challenge for engineering educators, since a lot of our domain research is responsible for creating this conundrum in the first place. As an example, 3-D printing has definitely made design and production more universal, but it has also made it more digital and easy to do with fewer and fewer people. We are at the cusp of producing more and more things for people who are employed less and less. What do we teach in this scenario and how do we impart knowledge whose half-life can be years or months or even weeks, and to what avail?
Aditya Johri is an associate professor in the Department of Information Sciences and Technology at George Mason University.