Rise of the Robots: Technology and the Threat of a Jobless Future (19 page)

One of the few economists to recognize offshoring’s disruptive potential is the former vice chairman of the Federal Reserve’s Board of Governors, Alan Blinder, who wrote a 2007 op-ed in the
Washington Post
entitled “Free Trade’s Great, but Offshoring Rattles Me.”
⁴⁸
Blinder has conducted a number of surveys aimed at assessing the future impact of offshoring and has estimated that 30–40 million US jobs—positions employing roughly a quarter of the workforce—are potentially offshorable. As he says, “We have so far barely seen the tip of the offshoring iceberg, the eventual dimensions of which may be staggering.”
⁴⁹

Virtually any occupation that primarily involves manipulating information and is not in some way anchored locally—for example, with a requirement for face-to-face interaction with customers—is potentially at risk from offshoring in the relatively near future and then from full automation somewhat further out. Full automation is simply the logical next step. As technology advances, we can expect that more and more of the routine tasks now performed by offshore workers will eventually be handled entirely by machines. This has already occurred with respect to some call center workers who have been replaced by voice automation technology. As truly powerful natural language systems like IBM’s Watson move into the customer service arena, huge numbers of offshore call center jobs are poised to be vaporized.

As this process unfolds, it seems likely that those companies—and nations—that have invested heavily in offshoring as a route to profitability and prosperity will have little choice but to move up the value chain. As more routine jobs are automated, higher-skill, professional jobs will be increasingly in the sights of the offshorers. One factor that is, I think, underappreciated is the extent to which advances in artificial intelligence as well as the big data revolution may act as a kind of catalyst, making a much broader range of high-skill jobs potentially offshorable. As we’ve seen, one of the tenets of the big data approach to management is that insights gleaned from algorithmic analysis can increasingly substitute for human judgment and experience. Even before advancing artificial intelligence applications reach the stage where full automation is possible, they will
become powerful tools that encapsulate ever more of the analytic intelligence and institutional knowledge that give a business its competitive advantage. A smart young offshore worker wielding such tools might soon be competitive with far more experienced professionals in developed countries who command very high salaries.

When offshoring is viewed in combination with automation, the potential aggregate impact on employment is staggering. In 2013, researchers at the University of Oxford’s Martin School conducted a detailed study of over seven hundred US job types and came to the conclusion that nearly 50 percent of jobs will ultimately be susceptible to full machine automation.
⁵⁰
Alan Blinder and Alan Krueger of Princeton University conducted a similar analysis with respect to offshoring and found that about 25 percent of US jobs are at risk of eventually being moved to low-wage countries.
⁵¹
Let’s hope there’s significant overlap between those two estimates! Indeed, in all likelihood there is plenty of overlap when the estimates are viewed in terms of job titles or descriptions. The story is different along the time dimension, however. Offshoring will often arrive first; to a significant degree, it will accelerate the impact of automation even as it drags higher-skill jobs into the threat zone.

As powerful AI-based tools make it easier for offshore workers to compete with their higher-paid counterparts in developed countries, advancing technology is also likely to upend many of our most basic assumptions about which types of jobs are potentially offshorable. Nearly everyone believes, for example, that occupations that require physical manipulation of the environment will always be safe. Yet, military pilots located in the western United States routinely operate drone aircraft in Afghanistan. By the same token, it is easy to envision remote-controlled machinery being operated by offshore workers who provide the visual perception and dexterity that, for the time being, continues to elude autonomous robots. A need for face-to-face interaction is another factor that is assumed to anchor a job locally. However, telepresence robots are pushing the frontier in
this area and have already been used to offshore English language instruction from Korean schools to the Philippines. In the not too distant future, advanced virtual reality environments will likewise make it even easier for workers to move seamlessly across national borders and engage directly with customers or clients.

As offshoring accelerates, college graduates in the United States and other advanced countries may face daunting competition based not just on wages but also on cognitive capability. The combined population of India and China amounts to roughly 2.6 billion people—or over eight times the population of the United States. The top 5 percent in terms of cognitive ability amounts to about 130 million people—or over 40 percent of the entire US population. In other words, the inescapable reality of the bell-curve distribution stipulates that there are far more very smart people in India and China than in the United States. That will not necessarily be a cause for concern, of course, as long as the domestic economies in those countries are capable of creating opportunities for all those smart workers. The evidence so far, however, suggests otherwise. India has built a major, nationally strategic industry specifically geared toward the electronic capture of American and European jobs. And China, even as the growth rate of its economy continues to be the envy of the world, struggles year after year to create sufficient white-collar jobs for its soaring population of new college graduates. In mid-2013, Chinese authorities acknowledged that only about half of the country’s current crop of college graduates had been able to find jobs, while more than 20 percent of the previous year’s graduates remained unemployed—and those figures are inflated when temporary and freelance work, as well as enrollment in graduate school and government-mandated “make work” positions, are regarded as full employment.
⁵²

Thus far, a lack of proficiency in English and other European languages has largely prevented skilled workers in China from competing aggressively in the offshoring industry. Once again, however, technology seems likely to eventually demolish this barrier.
Technologies like deep learning neural networks are poised to transport instantaneous machine voice translation from the realm of science fiction into the real world—and this could happen within the next few years. In June 2013, Hugo Barra, Google’s top Android executive, indicated that he expects a workable “universal translator” that could be used either in person or over the phone to be available within several years. Barra also noted that Google already has “near perfect” real-time voice translation between English and Portuguese.
⁵³
As more and more routine white-collar jobs fall to automation in countries throughout the world, it seems inevitable that competition will intensify to land one of the dwindling number of positions that remain beyond the reach of the machines. The very smartest people will have a significant advantage, and they won’t hesitate to look beyond national borders. In the absence of barriers to virtual immigration, the employment prospects for nonelite college-educated workers in developed economies could turn out to be pretty grim.

Education and Collaboration with the Machines

As technology advances and more jobs become susceptible to automation, the conventional solution has always been to offer workers more education and training so that they can step into to new, higher-skill roles. As we saw in
Chapter 1
, millions of lower-skill jobs in areas like fast food and retail are at risk as robots and self-service technologies begin to encroach aggressively in these areas. We can be sure that more education and training will be the primary proffered solution for these workers. Yet, the message of this chapter has been that the ongoing race between technology and education may well be approaching the endgame: the machines are coming for the higher-skill jobs as well.

Among economists who are tuned in to this trend, a new flavor of conventional wisdom is arising: the jobs of the future will involve
collaborating with the machines. Erik Brynjolfsson and Andrew McAfee of the Massachusetts Institute of Technology have been especially strong proponents of this idea, advising workers that they should learn to “race with the machines”—rather than against them.

While that may well be sage advice, it is nothing especially new. Learning to work with the prevailing technology has always been a good career strategy. We used to call it “learning computer skills.” Nevertheless, we should be very skeptical that this latest iteration will prove to be an adequate solution as information technology continues on its relentless exponential path.

The poster child for the machine-human symbiosis idea has come to be the relatively obscure game of freestyle chess. More than a decade after IBM’s Deep Blue computer defeated world chess champion Garry Kasparov, it is generally accepted that, in one-on-one contests between computers and humans, the machines now dominate absolutely. Freestyle chess, however, is a team sport. Groups of people, who are not necessarily world-class chess players individually, compete against each other and are allowed to freely consult with computer chess programs as they evaluate each move. As things stand in 2014, human teams with access to multiple chess algorithms are able to outmatch any single chess-playing computer.

There are a number of obvious problems with the idea that human-machine collaboration, rather than full automation, will come to dominate the workplaces of the future. The first is that the continued dominance of human-machine teams in freestyle chess is by no means assured. To me, the process that these teams use—evaluating and comparing the results from different chess algorithms before deciding on the best move—seems uncomfortably close to what IBM Watson does when it fires off hundreds of information-seeking algorithms and then succeeds in ranking the results. I don’t think it is much of a stretch to suggest that a “meta” chess-playing computer with access to multiple algorithms may ultimately defeat the human teams—especially if speed is an important factor.

Secondly, even if the human-machine team approach does offer an incremental advantage going forward, there is an important question as to whether employers will be willing to make the investment necessary to leverage that advantage. In spite of the mottos and slogans that corporations direct at their employees, the reality is that most businesses are not prepared to pay a significant premium for “world-class” performance when it comes to the bulk of the more routine work required in their operations. If you have any doubts about this, I’d suggest trying to call your cable company. Businesses
will
make the investment in areas that are critical to their core competency—in other words, the activities that give the business a competitive advantage. Again, this scenario is nothing new. And, more importantly, it doesn’t really involve any new people. The individuals that businesses are likely to hire and then couple with the best available technology are the same people who are largely immune to unemployment today. It is a small population of elite workers. Economist Tyler Cowen’s 2013 book
Average Is Over
quotes one freestyle chess insider who says that the very best players are “genetic freaks.”
⁵⁴
That hardly makes the machine collaboration idea sound like a systemic solution for masses of people pushed out of routine jobs. And, as we have just seen, there is also the problem of offshoring. A great many of those 2.6 billion people in India and China are going to be pretty eager to grab one of those elite jobs.

There are also good reasons to expect that many machine collaboration jobs will be relatively short-lived. Recall the example of WorkFusion and how the company’s machine learning algorithms incrementally automate the work performed by freelancers. The bottom line is that if you find yourself working with, or under the direction of, a smart software system, it’s probably a pretty good bet that—whether you’re aware of it or not—you are also training the software to ultimately replace you.

Yet another observation is that, in many cases, those workers who seek a machine collaboration job may well be in for a “be careful
what you wish for” epiphany. As one example, consider the current trends in legal discovery. When corporations engage in litigation, it becomes necessary to sift through enormous numbers of internal documents and decide which ones are potentially relevant to the case at hand. The rules require these to be provided to the opposing side, and there can be substantial legal penalties for failing to produce anything that might be pertinent. One of the paradoxes of the paperless office is that the sheer number of such documents, especially in the form of emails, has grown dramatically since the days of typewriters and paper. To deal with this overwhelming volume, law firms are employing new techniques.

The first approach involves full automation. So-called e-Discovery software is based on powerful algorithms that can analyze millions of electronic documents and automatically tease out the relevant ones. These algorithms go far beyond simple key-word searches and often incorporate machine learning techniques that can isolate relevant concepts even when specific phrases are not present.
⁵⁵
One direct result has been the evaporation of large numbers of jobs for lawyers and paralegals who once would have sorted laboriously through cardboard boxes full of paper documents.

There is also a second approach in common use: law firms may outsource this discovery work to specialists who hire legions of recent law school graduates. These graduates are typically victims of the bursting law school enrollment bubble. Unable to find employment as full-fledged lawyers—and often burdened with enormous student loans—they instead work as document reviewers. Each attorney sits in front of a monitor where a continuous stream of documents is displayed. Along with the document, there are two buttons: “Relevant” and “Not Relevant.” The law school graduates scan the document on the screen and click the proper button. A new document then appears.
⁵⁶
They may be expected to categorize up to eighty documents per hour.
⁵⁷
For these young attorneys, there are no courtrooms, no opportunity to learn or to grow in their profession, and no
opportunity for advancement. Instead, there are—hour after hour—the “Relevant” and “Not Relevant” buttons.
*