There has been a welter of articles in recent months about the profound impact robotics and artificial intelligence (AI) will have on employment in the coming years. I have been reading extensively in the research behind these reports, and I am convinced that the conclusions are well founded. This is not a “news cycle fad”; the nature of work is on the cusp of changing in astonishing ways.
One helpful way to think about technological obsolescence is to envision a set of four quadrants in which the top left is comprised of highly skilled functions that are routine; the bottom left, unskilled or semi-skilled routine jobs; bottom right, lower-skilled non-routine work; and upper right, highly-skilled non-routine work.
For the past two centuries, throughout the industrial revolution, jobs in the bottom left have been the most vulnerable to replacement by machines. In textile manufacturing, steel production, agriculture, and many other industries, technology ranging from the steam engine and the McCormick reaper to Amazon’s roboticized order-fulfillment centers have for centuries replaced lower-skilled jobs whose routine nature rendered them vulnerable to machinery that was well-suited to repetitive performance.
At the dawn of the computer and robotics revolutions, most people thought that the bottom right quadrant would be the next to fall: lower-skilled non-routine jobs. Yet it didn’t turn out that way. As it happens, automation is less concerned with the skill set required for a job so much as the routine repetitiveness of a job. Even highly skilled routine jobs are easier to automate than the lowest-skilled non-routine task. In other words, it is easier to build software that analyzes financial data—high-skilled, routine work—than to walk about in an unfamiliar environment—a low-skilled, non-routine task that even a two-year-old human can do.
For employees, this means that the most immediate jobs and careers under threat of being rendered obsolete in our era are all those based on routine tasks in the left-hand quadrants, skilled or otherwise. I suspect, for instance, that pharmacists, who, despite their high degree of training, perform an essentially repetitive task, will soon be replaced by robots that don’t require seven years of education in pharmaceutical colleges, don’t expect benefits, and make fewer mistakes filling prescriptions. To take another example, even many of the more routine news stories you read today (such as the daily stock market reports) are written by journalistic bots.
Importantly, though, we are seeing tremendous progress on the non-routine side of our quadrants, which will prove even more disruptive over the next quarter century. Robots are getting much better at human-like conversational interaction, at translating colloquially from one language to another, and at driving cars in ever-shifting traffic conditions, to take but a few examples from just the past week’s news stories. Astonishingly, too, software bots are increasingly writing legal contracts or generating medical diagnoses. (IBM’s Watson, the Jeopardy champion, is now being dedicated to cancer diagnostics; Watson has already absorbed the content of 750,000 medical journal articles on cancer to prepare for the task.)
A 2013 Oxford University study concluded that 47 percent of American jobs will be eliminated in the next 20 years. Think about that. Half of all jobs gone in two decades. Always, in the past, new jobs have been created by technological innovation at a rate that has roughly kept pace with the number of jobs eliminated, ensuring that workers (with retraining) could find replacement work. Yet it is hard to imagine new job creation keeping up with today’s scale and rapidity of worker displacement.
What does all this mean for education in 2017? I would argue that we can no longer assume that we are preparing students to work in knowledge economy jobs that historically have been immune to automation. Even doctors and lawyers will increasingly be replaced by bots programmed with ever-deeper AI. (And remember: As long as Moore’s Law pertains, the chips running bots will get twice as powerful every 18 months at the same time they become 30 percent less expensive to make each year, while humans, evolving at an infinitesimally smaller pace, show no discernible increase in their cranial capacity every 18 months.) The displacement of even the most highly educated “professionals” is but a short time horizon away.
What, then, will remain uniquely human domains in the world of work? This is what Greg Toppo (National Education & Demographics Writer at USA Today and author of The Game Believes in You) and I are in the process of co-authoring a book about. [Click here to read an earlier conversation with Greg]. Yet I would venture preliminarily here that creativity, entrepreneurship, goal-setting and large-scale strategic decision-making will almost certainly remain exclusively within human purview for the foreseeable future. Also, areas pertaining to nurturance, caring, and deep social values are unlikely to be replaced by machines. (People in hospice care will want to hold another person’s hand, not a robot’s.) Also, humans will remain the strategic goal setters. (After all, only people can decide the purposes toward which AI should be applied.)
This calls for a dramatic reconsideration of the “knowledge economy” careers for which our schools continue preparing students. Why educate students for jobs that won’t exist? Perhaps we should place far greater focus on teaching children to be innovative, entrepreneurial creators and strategic managers, as well as to be more nurturing and, well, more fully human.
Jim Tracy is Head of School at Rocky Hill School in East Greenwich, RI.