On my first day of class in fifth grade, my teacher told us not to worry about school because robots would be doing everything by the time we became adults. I liked her immediately. Recently, there’s been speculation by notable technology entrepreneurs that the development of fully functioning robots and the “coming of age” of artificial intelligence will mean fewer jobs ahead. No one, including me, can predict the future, so this article is as speculative as any other on the topic. But there is little if any historical precedent indicating that the development of technology has decreased overall employment since the advent of the Industrial Revolution.
In 1867, Karl Marx published Das Kapital. Marx and others were rightfully concerned about the use of machines and how that would affect labor. On the surface, this makes total sense. If an automated drill press can drill 400 holes per hour and a person can drill only 40 holes per hour, and we have to produce 3,200 holes per day, then just one drill press can do the work of 10 people. Marx claimed that the development of machines would make people less necessary in producing goods. What really happened is that the nature of employment changed. By 1910, here in America, only 31 percent of laborers were employed in agriculture, while 100 years earlier, approximately 81 percent were tied to the farm.
Today, leading technologists like Bill Gates, Mark Cuban, and Elon Musk seem to be raising new alarms over job losses tied to automation. I believe they will be wrong. In our country, if you are famous or rich, it seems to give you a license on being an expert in many other things such as politics or history or, say, predicting the future. In my preceding example, to end the analysis by saying only one drill press needed is too simple and static. Even if one drill press replaces 10 hole-drillers, the economy will also probably need people in a factory to make drill presses. There would be a shift of work from people who drill holes to people who make drill presses. It’s also likely that drill presses require steel, which has to be produced. We have seen this dynamic throughout history. Henry Ford’s development of the Model T automobile, for example, led to the need for fewer blacksmiths but more people to build roads, tires, and toll booths.
Of course, another important implication in the argument today is that robots somehow represent either a greater scope of change or a different type of change than did industrial machines. This is not surprising; we tend to applaud our technology increases without respect to what happened before us. Personally, I think the move from people using a plow to using a combine harvester in the 19th century is way more significant on a relative basis than the development of robots today. The relative rate of technology change over the past 50 years is much less than it was in the roughly 50 years following the end of the Civil War, for example. It seems to me that making a car drive itself is much less complicated and innovative than was inventing the first car. Prior to the 19th century, innovation had been existent, but it was incremental. There wasn’t much difference between the way people lived in the medieval period and the way they lived in the late 18th century. But can you imagine someone born in 1865 who lived the first 25 years of his life and then, a la Rip Van Winkle, slept for 20 years and then awoke in 1910 to see cars, planes, electric lights, movie cameras, and phonographs?
To the specific points made in the current debate, some assume that these technological changes will occur in a vacuum; they assume that robots won’t create a need for something else. I’m not even going to address the fact that artificial intelligence has been over-promised for at least the past 25 years. But let’s assume for argument’s sake that robots are going to replace a substantial amount of the physical work that people do today. Just like the need for drill press factories in the above example, there will still likely be robot-building facilities, since manufacturing is involved. If there’s artificial intelligence involved, there will probably be a need for software development. In addition, typically in our system, initially large profits tend to beget additional suppliers chasing those profits. As new suppliers emerge, there is a need for more labor. Finally and most important, as we’ve seen in the past 40 years, technology seems to generate even more technology and more demand for “stuff.” When I was young, every family in my neighborhood was lucky to have even one television set. Now, people have three and four sets or ginormous ones, and households have gotten used to having multiple cars and computers, despite the fact that cars and computers have become more efficient and powerful and can do more. In other words, technology advances create an increase in the standard of living (I suppose more Kardashian episodes is a better standard of living), which tends to lead people to consume more things, and that consumption leads to a need for additional production across the board. This doesn’t even take into account what the producers of robots might do with all of their savings and new profits. Will they create more new technology or new businesses? Probably.
I realize this is more or less the standard capitalist argument, but despite some periods of underemployment, there’s been no relationship between the development of technology and unemployment to date, at least over the past 100 years. Nobody knows for certain what’s going to happen in the future, but technology leaders and fifth grade teachers alike tend to look at the future through the prism of their own worlds rather than through the broader view of objective history.