AI is Not a Job Killer — it’s a Job Category Killer
AI brings mixed emotions and opinions when referenced in the context of jobs. Here’s something fun you can do at your next business or technology networking event; ask the question “Do you think Artificial Intelligence will be a net job killer or net job creator?” What we would expect you would hear is an interesting and conflicting set of opinions that range from “AI will destroy all jobs as we know it” to “AI will enable us to work better and do new things we’ve never been able to do.” Do a quick Google search on the topic and almost every other link is something contradicting the link before it. If you look at various economic and analyst predictions, their assessments are all over the place, ranging from dramatic job losses across most economic sectors to large increases in employment due to dramatic increases in job productivity. Of course, as with everything, the true answer will be somewhere in the middle. There’s no doubt that AI will eliminate the need for many different kinds of jobs in many different categories. But at the same time, AI will create new jobs in categories that we know of, and many more in categories that have yet to be created.
AI Will Destroy Jobs
According to a Gallup poll, many people think AI will destroy jobs, but just not theirs. In fact, in that poll, over 73% of Americans believe that AI will be a net job destroyer, but only 23% of these same surveyed adults were worried about it. How can you reconcile those two positions? In that same survey, over 90% of those polled believed that AI will destroy at least half of all jobs, but 91% of them believed it wouldn’t impact their employment. The general consensus from the average person is that AI is continuing the unimpeded progress of automation and technology that brings increasing levels of productivity. By productivity, we mean it in the economic sense in that businesses are able to do more with their existing resources, and perhaps do more with fewer resources, with those resources being primarily human labor and the costs associated with it.
We’ve discussed many times that automation is not intelligence, and there is no doubt that the addition of more cognitive capabilities in the technologies that companies use will allow organizations to rethink their usage of human labor for activities as wide ranging as call center operations, warehouse activities, trucking and transportation, brick-and-mortar retail, and even mining, oil, and gas activities. Many make the argument that job losses won’t be felt at the top or bottoms of the pay scale, but rather in the middle. There will be less management by humans, and more management by machines. Collectively these middle income jobs employ a very large percentage of the American population. According to the 2017 Bureau of Labor & Statistics report, most employment is in retail, professional services, healthcare, and government.
For sure, we can expect AI to take a chunk out of retail, government, and professional services employment. We’re already experiencing a conversion of previously human-intensive labor that involved moving paper or bits and bytes around from one place to another going away as systems become increasingly more integrated and cognitive services replace humans at highly repetitive, regulatory intensive, and error-prone processes. Why have a human move information around when a computer can do just as good a job, especially with the ability to understand the meaning and context of information? Furthermore, while we’re still in the very early (and possibly dangerous) days of autonomous vehicles, there’s no question that the future direction of the transportation, warehouse, and logistics industries is rapidly heading to an autonomous future. Truck driving, which represents the largest employer by number of people employed by category in many US states will clearly be a job position in jeopardy in the future.
No, Wait, AI Will Create Jobs
The argument with every wave of technology, from the automatic weaving looms of the early industrial revolution to the computers of today is that jobs are not destroyed, but rather employment shifts from one place to another as entirely new categories of employment are created. The Luddites might have wrecked the mills as a protest against machine-enabled automation, but today, those same workers would be defending manufacturing against the disappearance of those jobs. In fact, if you zoom out far enough, you’ll see the trend away from manufacturing and towards professional services started just after World War II and the irresistible move towards globalization. Some have even made the stark point that the only reason why American manufacturing stayed a significant employer until the late 1970s was because World War II destroyed so much manufacturing capacity in Europe and Asia that the United States was the de facto manufacturing leader for many decades afterwards.
Let’s take another look at the BLS labor & statistics report and compare that to previous years. In 1910, manufacturing, transportation, retail, and domestic services were the major employers. Inventions such as the washing machine, dishwasher, microwave, and cooktops / ranges put an end to domestic service as a major employer. Yet, the US didn’t experience massive waves of unemployment, because we invented whole new major sectors of the economy in professional services which barely even existed as a category in 1910. Likewise, the evolution of major employer by category shifted significantly from even 1978, when secretaries were the largest category of employee type. Along came computers and out went entire rooms of typewriters, filing cabinets, and people handling the scheduling of management staff. Yet, again, the United States didn’t have a major economic collapse or massive waves of unemployment because new employment categories emerged before the old ones were fully retired.
Furthermore, there are many categories of employment that will be relatively un-impacted by even super-cognitive AI-enabled systems. The government, healthcare, education, leisure and hospitality, and many professional services categories (notably real estate) will continue being major employers. And we’ve made the point that the addition of AI to these industries will just make humans better at their jobs and more responsive to the needs of other humans, rather than replacing them. After all there’s nothing more human than government or lounging in a hotel. As long as there are humans, we’ll have government, teachers, doctors, and cabanas. Even in areas where robotics has had a significant impact, we see puzzling contradictions. Amazon is employing more people than ever in its warehouses even though it has one of the highest penetrations of intelligent automation.
AI Will Kill Jobs in Some Categories, but Create Entirely New Categories of Jobs
However, this is missing the huge point. The real job creator will be entirely new categories of jobs that we can’t even imagine today. If the “2018 you” went back in time to talk to “1988 you” and you told yourself that you’ve got a job in social media marketing, you’d blankly stare back at the future you and wonder what the heck you were talking about. Likewise, the 2048 (really old) you would tell yourself about whole sectors of the economy and major employers that would not even be possible today. Yesterday’s manufacturers are today’s programmers. Yesterday’s secretaries are today’s database administrators. Yesterday’s milkmen are today’s Uber drivers. Indeed, it’s not that jobs have been created or destroyed, but rather entire job categories are gone and new ones have taken their place.
Humans have a very, very hard time imagining what sort of new jobs and new sectors will come to be. Since we can only truthfully envision what might be gone tomorrow, but we can’t imagine what will be possible, our limited human minds make the mental calculation that the net impact must be negative. But that’s simply not the case. For sure there will be new categories of jobs that relate directly to creation of AI-enabled capabilities, whether in software or hardware. However, no one can argue that any number of those jobs will make up for the decline in workers that come from using those products. Data scientists, robotic engineers, and ML programmers will not make up for fewer truck drivers or call center workers.
Rather, the creation will be where we least expect it. Will there be entirely new categories of professional services that would be impossible without the use of advanced intelligence as a helper? Will we see a massive wave of self-employment enabled by intelligent assistants and third-party services that turn individuals into major powerhouse corporations of one?
There is no question that the world’s economies are undergoing a fundamental, tectonic shift as we move from one age of industrialization to another — something that many in the industry are calling Industry 4.0. The transition between each wave of industrialization is not necessarily a clean one. If new job categories are not created before old job categories are retired, then the transition can be a messy one. However, it’s clear that we’re already embarked on the transition, and now it remains to be seen what the future will create with the new capabilities that are being developed.