Robots are coming for our jobs. That’s what we’ve been told. And it’s true, to an extent.
I could cite any number of these doomsday scenarios. I’ll choose this one from David Ignatius at The Washington Post:
The deeper problem facing the United States is how to provide meaningful work and good wages for the tens of millions of truck drivers, accountants, factory workers and office clerks whose jobs will disappear in coming years because of robots, driverless vehicles and “machine learning” systems.
Scary, right? That’s one downside of technology. We get so good at automating so much work, that we could potentially wipe out whole classes of jobs.
Our fear of job loss from automation isn’t exactly new
For better or worse, we have feared exactly this outcome for at least 100 years. I’ll quote at length from Henry Ford’s autobiography, My Life and Work, which was published in 1922:
But how about production? If every necessity of life were produced so cheaply and in such quantities, would not the world shortly be surfeited with goods? Will there not come a point when, regardless of price, people simply will not want anything more than what they already have? And if in the process of manufacturing fewer and fewer men are used, what is going to become of these men–how are they going to find jobs and live?
Take the second point first. We mentioned many machines and many methods that displaced great numbers of men and then someone asks:
“Yes, that is a very fine idea from the standpoint of the proprietor, but how about these poor fellows whose jobs are taken away from them?”
The question is entirely reasonable, but it is a little curious that it should be asked. For when were men ever really put out of work by the bettering of industrial processes? The stage-coach drivers lost their jobs with the coming of the railways. Should we have prohibited the railways and kept the stage-coach drivers? Were there more men working with the stage-coaches than are working on the railways? Should we have prevented the taxicab because its coming took the bread out of the mouths of the horse-cab drivers? How does the number of taxicabs compare with the number of horse-cabs when the latter were in their prime? The coming of shoe machinery closed most of the shops of those who made shoes by hand. When shoes were made by hand, only the very well-to-do could own more than a single pair of shoes, and most working people went barefooted in summer. Now, hardly any one has only one pair of shoes, and shoe making is a great industry. No, every time you can so arrange that one man will do the work of two, you so add to the wealth of the country that there will be a new and better job for the man who is displaced. If whole industries changed overnight, then disposing of the surplus men would be a problem, but these changes do not occur as rapidly as that. They come gradually. In our own experience a new place always opens for a man a soon as better processes have taken his old job. And what happens in my shops happens everywhere in industry. There are many times more men to-day employed in the steel industries than there were in the days when every operation was by hand. It has to be so. It always is so and always will be so. And if any man cannot see it, it is because he will not look beyond his own nose.
In all seriousness, Mr. Ford is making the right point. Unfortunately, it’s a difficult point to sympathize with, since good people face hard times as a result of automation.
What I found interesting about Henry Ford’s point is that he goes deeper than most automation advocates. The standard pro-automation argument is that we gain higher quality jobs than we lose. So, yes, truck drivers lose their jobs. But we gain high quality jobs for engineers that will design and optimize the routing software for the driverless fleets.
Mr. Ford makes a more ambitious point. He says that automation actually fuels job growth in the fields it disrupts.
A way automation can actually promote job growth
How would this work for truck drivers? The idea is that driverless trucking would drastically reduce the cost of transporting certain goods across certain distances. The costs would fall so much, that the demand for trucking services would skyrocket. So, yes, we would lose jobs for truckers. But we would gain jobs for all the people involved in truck maintenance, or in managing fleet logistics, or in designing fuel efficiency improvements.
Let’s zoom in on the fuel efficiency angle for a moment. Obviously trucking companies care about fuel efficiency today. But what happens if driverless trucking drastically reduces transportation costs, which then spurs massive demand growth? Trucking, as an industry, has much greater scale. As any business grows, small improvements matter more and more, as they’re replicated across a larger system.
Today, trucking companies might be able to justify technologies that yield at least 5% improvement in fuel efficiency. But if the industry is much larger, and costs are much lower, then the industry might be able to justify efforts aimed at 1% improvement. Whole areas of work that are uneconomic today become economic tomorrow, in a new low cost, high volume reality.
Further, it doesn’t matter whether we’re using hydrocarbon fuels or renewable energy for our fictitious fleets. Every trucking operator will want to know how to travel further on one unit of energy. Maybe you have to design better internal combustion engines. Maybe you have to design better batteries for electric cars. Who knows. The point is, the winning technology doesn’t have to take any particular form. Any reality where trucking demand ramps up leads to job growth.
That’s a point I didn’t appreciate until I read Henry Ford’s take on this issue. I was content with the idea of high quality jobs replacing low quality jobs. But reality is even more favorable than that. It’s funny to read about automated shoe manufacturing replacing manual assembly, but it was the reality of Ford’s time. These kinds of examples keep appearing throughout history.
The Lindy effect and automation
That brings me to one last point. I’m reading Antifragile by Nassim Taleb at the moment. Taleb talks about the Lindy effect, which speaks to the lifetime of nonperishable things like books or ideas. For perishable items, like living things, each day they’re alive effectively reduces the number of days they will remain alive in the future. For nonperishable items, the Lindy effect says the opposite is true. For each day an idea is alive, we should expect it to live even longer into the future.
Henry Ford’s idea about the benevolence of automation is 100 years old. For 100 years, his idea has borne fruit. The Lindy effect tells us we should expect this idea to continue to serve us well going forward.
Automation is a powerful force. It’s largely, though not entirely, benevolent. It has served us and our economy well.
We shouldn’t fear automation. We do need to ensure we offer new paths for displaced workers. As a society, we need to do a good job adapting smoothly to the changes brought by automation.
But we shouldn’t fear these changes. Automation makes us wealthier. It offers us better jobs. We just need to be clever enough to take proper advantage of it.