I’ve been concerned about the impact of automation on human employment for a while. My research activity started after November 2016, when I decided I needed to learn whether the much-touted phrase, “We’re bringing back the jobs,” was valid or not.

While it is true that some jobs in the United States have been exported to other countries, a study coming out of Ball State University in Indiana asserts that 88 percent of the jobs lost by U.S. workers is due to automation. Combine this study with the one published in 2013 by Frey and Osborne out of Oxford University that says that 47 percent of the jobs in the United States are risk, then things don’t look so rosy. Martin Ford, in his book, “Rise of the Robots,” goes so far as to say that every occupation is at risk, not just low wage workers who are paid to do repetitive tasks that are easy to learn. Not only will we have machines flipping burgers, they’ll also be perusing volumes of legal documents during discovery processes, which is the type of work that now is done by paralegals and newly hired law school graduates.

Machines are getting smarter all the time and they are getting cheaper. You can buy Baxter, a full fledged factory robot, for $22,000. Baxter can work 24/7 and is not subject to OSHA laws or workers’ compensation coverage.

Software? That’s getting cheaper, too. Every day, another open-source project is made available on GitHub that allows anybody with a computer and connection to the internet to take advantage of the power of artificial intelligence and machine learning. You want Swift code for artificial intelligence? No problem. It’s on GitHub and available to all for free! Want to convert your English ReadMe into Japanese? Simple, just use an API such as Microsoft Translator or Google Cloud Translate.

Given the current trends in technology, it’s not hard to imagine that within the next 15 years to 20 years we’re going to have a world in which fewer people are employed. This future scenario worries me.

About a week ago I was sharing my concerns with Chris Surdak at Surdak and Company. Chris works with large corporations implementing robotic process automation (RPA). He is also on the board of the Institute for Robotic Process Automation and Artificial Intelligence. Chris is at the front lines of using robotics and artificial intelligence to increase commercial efficiency. He and I share a concern about the wave of worker displacement we see on the employment horizon. However, Chris has an interesting take: While he agrees that automation is going to cause significant displace of human workers, he sees an obstacle.

Imagine you are an executive in a corporation and you have 5000 people under you. One day someone like me tells you that you can replace 80 percent of your workers, 4000 people, by streamlining your work processes using robotics and artificial intelligence. You should be ecstatic, right? That is a significant cost savings that is going to directly affect your bottom line.

But, it doesn’t work that way because that is not how many corporate managers think. They want to keep their jobs and in order to do so they need two things: headcount and budget. So, if I come along and tell them that I can create an efficiency that is effectively going to cost them power (less employees and a smaller budget to manage) and maybe their jobs, do you think they are going to jump on board right away? Nope. In fact, they are going to do everything that can to slow down my work until they retire.

Chris’s statement was a revelation to me. Previously, I thought corporations in their continual quest to seek greater profits will do whatever is required to reduce costs and increase efficiency. Most of us who do tech, particularly in DevOps, have embraced this notion. Automation is our friend. We’re lazy. We want the code and machines to do the heavy lifting. And, we’ve created environment where automation is necessary, For example, large-scale container deployment alone went beyond the capabilities of human management a long time ago.

There is a good argument to be made that the factory floor has not had problems using machine automation to improve efficiency. Factories have looked for improved efficiency via tooling since before industrial textile mills used water wheels to power their looms. But, there are other industries that are risk-averse and slow to change; insurance, for instance. The virtue of an industry such as insurance is that it’s ingrained in its culture to manage risk and to be measured when considering change. After all, would you want your retirement annuity to be subject to the whims of a day trader? No. You want your money well-protected by a cautious steward. However, with such safety comes inertia. Thus, there is an intrinsic resistance to change in general and automation in particular. Add in a culture of self preservation on the part of management and … well, you get the picture.

So there are industries that will be resistant—for a while, anyway. Then one of two things will happen. The companies that are not resistant to automation will enjoy the efficiencies the technologies bring. They will gain market share by providing a better product at lower cost and they will trample out the competition—think Uber and the taxi industry. Or, those resistant corporations will realize that to survive, managers who manage by “saving my job” will be given a golden parachute after revealing all the tribal knowledge hidden in their work groups. (Undocumented rules and processes are the bane of process engineers when implementing automation.) Either way, the march of automation will continue and workers will be displaced.

Given the increasing rate at which jobs are being replaced by automation, will these displaced workers find other jobs? This question I leave to you to answer.

— Bob Reselman