A great deal of media coverage and a number of recent books warn about the threat to jobs from a surging wave of technological innovation that is leading to such breakthroughs in automation as self-driving cars and robot home assistants. We are in the early days of what Massachusetts Institute of Technology (MIT) economists Erik Brynjolfsson and Andrew McAfee dubbed the "Second Machine Age" in their influential 2014 book of the same title. This line of argument suggests that the skills that will secure people gainful employment in this emerging era are those learned by an education in the STEM fields—science, technology, engineering, and math. Earning a liberal arts degree is characterized as an impractical extravagance the workers of the future simply can't afford.
Reading novels and poems, revisiting the debates of ancient philosophy, or studying the history of the French Revolution or the culture of a remote island community isn't likely to get you a decent job in today's more tech- driven economy, and certainly not in the future, or so the argument goes. Microsoft founder Bill Gates caused a stir in a speech to the National Governors Association that state funding in support of liberal arts education should be cut and more money dedicated to higher education in the STEM fields because those are the skills that will get people well-paying jobs. Billionaire cofounder of Sun Microsystems Vinod Khosla, now a leading venture capitalist investing in technology startups, has gone so far as to say that "little of the material taught in liberal arts programs today is relevant to the future." Silicon Valley venture capitalist and software pioneer who created the search engine Netscape Marc Andreessen quipped that those who learn the "soft skills" of the liberal arts in college, rather than the "hard skills" of science and tech, "likely will end up working in shoe stores."
Be Concerned, Not Fearful
The alarm about the future of work, and about the prospects of liberal arts graduates, is clearly founded in genuine concern, but it is also sorely misguided for a number of reasons. First of all, as we'll explore more fully in Chapter 8, though increasingly "smart" and nimble machines will likely be taking the place of some workers, the extent of likely job displacement has been greatly exaggerated. The threat to some jobs is not only clear; it is already present. Robots will assume more and more tasks that can be fully automated, as they've already done with jobs on the assembly lines of so many manufacturers. But the proportion of jobs that can be fully automated is more limited than suggested by forecasts. In many jobs, a number of tasks that can be automated, because they are routine or can be better performed by crunching vast amounts of data, will be taken over by machines. But in many cases, the result will not be to displace human workers; rather it will be liberate people to spend more time on the aspects of their work that require uniquely human skills—nonroutine tasks and complex problem solving that machines can't perform and may never be able to do.
Look no further than the legal profession to see change in the making. In 2015, MIT labor economist Frank Levy coauthored a paper with Dana Remus of the University of North Carolina School of Law called "Can Robots Be Lawyers? Computers, Lawyers, and the Practice of Law." The paper examined the notion that the legal profession is susceptible to automation, and that lawyers will soon be replaced by computers. Their argument was inspired by the advent of software designed to read and analyze legal documents during the discovery process.
After an extensive analysis of the time spent on individual tasks performed by attorneys, Levy and Remus found that they spend the majority of their time analyzing documents, counseling clients, and appearing in court, and many of the skills that make a legal professional especially effective, such as being able to think on their feet and interacting with clients, are, and will remain, uniquely human. They estimate that around 13 percent of legal work could one day be automated—a measurable amount, but relatively modest, especially since the change will happen over the course of many years. Instead of replacing lawyers, automation software will make lawyers more efficient. Machines will handle routine tasks; lawyers will do the rest.
A great irony of the discussion of job displacement is that among the jobs vulnerable to dislocation and automation are many in computer programming—currently hailed as both higher paying and in the highest demand. How might this happen? First, many of these jobs will be transferred overseas to developing economies invested in training masses of highly qualified - programmers—places like India, China, and Nigeria. These programmers are no longer simply an inexpensive workforce hired just to do the relatively simple work of building websites; they are being trained to great proficiency. Andela, a startup that aims to train one hundred thousand African programmers over the next decade, has so many applicants that its acceptance rate is under 1 percent. Andela invests as much as $10,000 to train each fellow in the latest software development, fellows such as Olajumoke Oladimeji, a young woman who already had a degree in computer science and electrical engineering from Lagos State University. The fellows are then matched with global companies. Due to the high pay programmers can demand, sending a good deal of programming work overseas is as unavoidable as was sending manufacturing to the developing world. In 1970 one in four Americans worked in manufacturing, whereas today it's fewer than one in ten. The flight of routine computing work will likely follow a similar pattern. Technical skill is important, but a technical education on its own will not automatically ensure employment in the Second Machine Age.
That said, the value of getting a high-quality STEM education, not one limited to learning computer programming languages, but a rigorous grounding in one of the hard sciences or fields of engineering, should not be contested. Work in pure scientific research, as well as in R&D in industry and at the high end of technological innovation, will likely always be secure. Regarding computer programming jobs, there is a current shortfall of candidates to fill them in the U.S. labor force, and the pace of change is rapid. In addition, analysis of the future needs of the job market has indicated that this shortfall will worsen significantly in the coming years. The Bureau of Labor Statistics approximates that there will be one million more computer science job openings by 2020 than domestic candidates qualified to fill them. This is a driving force behind the calls for more STEM majors, and we can't deny that there is a very real need for more techies.