The press has recently been awash in stories of worker robots of ever-increasing ability making human labor obsolete. Is it true that we are approaching a future of a mechanized workforce where humans need not apply? Experts’ opinions on the future of technological development vary spectacularly, as do views on the nature and magnitude of robotics’ impact on society and mankind’s ability to adapt. But one thing is collectively agreed upon: robotics has begun and will continue to transform the workforce in profound and inevitable ways.
Fears that automation will lead to unemployment have existed for centuries, but humans have shown a remarkable ability to adapt to modernization of work. While the percentage of Americans that plowed the fields dwindled from 33 percent to two percent over the last century, for instance, countless unforeseen occupations materialized. Some people believe the present situation to be no different: if robots eventually do some kinds of labor more efficiently than humans, people will simply move on to other work. As University of Pennsylvania School of Engineering and Applied Science dean Vijay Kumar told the HPR, “The implication that jobs will disappear and not be replaced is, I think, completely false.”
What makes this round of innovation potentially different is that robots have become smart. “Computers and other digital advances are doing for mental power … what the steam engine and its descendants did for muscle power,” write MIT Initiative on the Digital Economy co-directors Andrew McAfee and Erik Brynjolfsson in The Second Machine Age. A 2013 paper by Oxford academics Carl B. Frey and Michael A. Osborne predicts that 47 percent of U.S. jobs are at risk of automation.
Smart technologies already exist that demonstrate the vulnerability of white-collar service jobs to automation. There are computers that can make medical diagnoses with fewer errors than humans, and one venture capital firm has given a computer algorithm a voting position on its board of directors. Creative jobs are not immune, either. There are song-writing robots whose compositions are virtually indistinguishable from humans’, and, according to a March 7 op-ed in the New York Times by Shelley Podolny, robots are responsible for writing a “shocking” amount of news stories. All prompt the question: how will these technologies affect human employability?
Forecasting Machine Intelligence
Experts disagree immensely in both their outlook for future advances and their perception of the aptitude of present-day robots. On the one hand, some believe that technology will never be able to fully compete with the human mind. MIT economics professor David Autor told the HPR that the short- and medium-term consequences of robotics are over-hyped, because progress is “relatively close to a standstill. … Moreover, robotics and machinery are still extremely limited in terms of flexibility, adaptability, autonomy, and the ability to make independent decisions.”
Autor cited recent stumbles at the DARPA Robotics Challenge 2015, which required robots to compete tasks that would have allowed them to enter nuclear reactors and prevent the 2011 Fukushima nuclear disaster. “These bots were incredibly slow and incredibly fragile. There’s a viral video of one falling down crashing,” he recalled. “All these headlines and articles saying robots are going to take your job tomorrow are ridiculous.”
On the other side of the spectrum, a group of thinkers believes that it is simply a matter of time before robots surpass the intellectual capacity of humans—some even predict that this will happen in the foreseeable future. Proponents of the “technological singularity,” like futurist Ray Kurzweil in his book The Singularity is Near, believe that humans will design a machine smarter than themselves and launch a “cycle of machine intelligence’s iteratively improving its own design,” leaving mankind in the dust. Based on the reasoning that technological progress has been increasing at an exponential rather than linear rate, some in this school expect this event to unfold within this century.
The theory is highly controversial, but it has the support of a substantial number of artificial intelligence (AI) researchers. Median estimates in a 2012-2013 survey of around 550 AI experts were that high-level machine intelligence will develop in the 2040s with a 50 percent chance, and that a superintelligence, or “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest,” would develop 30 years thereafter. Perhaps it would be an anthropocentric fallacy to set human intelligence as an upper limit.
While it is easy to become mired in dissension about the future of technology, it is more illuminating to consider the different possible effects on the human workforce over the range of technological prospects. To begin, consider scenarios where people maintain distinct advantages over robotics and employment remains high, a situation particularly germane to technological skeptics and near-future predictions.
The Human Niche of the Future
Even the most conservative projections see robots playing a growing part in the workforce and replacing people in many important roles. Under certain forecasts, however, humans will continue to be employed, albeit in novel ways. Those like Harvard Kennedy School science and technology studies professor Sheila Jasanoff expect the robotics industry itself to engender new forms of occupation. “If you introduce revolutionary new technologies, you change the entire system. These robots won’t maintain themselves, right?” she remarked in an interview with the HPR. “When personal computing devices were introduced, people didn’t imagine that the app industry would take off.”
Human employability optimists also tend to view robotics as powerful but fundamentally inadequate tools that aid rather than replace humans. “Humans are better than robots at abstraction, generalization, and creative thinking” and robots can only solve structured problems in familiar environments, MIT Computer Science and Artificial Intelligence Laboratory director Daniela Rus wrote in a June article in Foreign Affairs. Instead of the singularity, many believe in what University of California, Berkeley professor Ken Goldberg dubs a “multiplicity,” where humans and machines work together in integrated teams, each complementing the skills of the other. This is the vision held by some industry leaders like InTouch Health founder Yulun Wang, who told HPR that his company’s medical robots are meant not to supersede medical staff but to “extend the reach of physicians and nurses.”
Some trust that even if the technical deficiencies of robots are overcome, humans will remain indispensable due to the importance of their unique qualities. According to the University of Pennsylvania’s Kumar, only humans can be counted upon for navigating the ethical decisions constantly required of us: “Imagine you are driving a car and see a big truck and a mother with a baby stroller. A self-driving car will think about self-preservation and will run over the stroller. You would make the moral judgment and most likely hit the truck.”
The desire for a human voice or perceived empathy could lead humans to price human receptionists over automated ones. And the part of our brain that lights up at the phrase “authentic handcrafted” may perpetually guarantee a market for the human brand—even if we manage to churn out AI capable of perfect emotional intelligence, humans may prefer having a warm-blooded personal trainer to one of plastic and metal.
When Humans Workers are Superfluous
At the other side of the debate are experts who envision a future where the number of employable humans reduces substantially. Their more pessimistic predictions stem from a belief that current automation technologies will become more widespread, cheaper, and easier to integrate into the workforce. They may also doubt that people can always find replacement jobs for those lost to technology, or that humans will maintain a competitive advantage against AI.
Many question the optimists who say that as many new jobs will be gained as will be lost from automation. In a Wall Street Journal essay from July 2014, former Secretary of the Treasury and Harvard University president Lawrence Summers warned of lasting structural unemployment due to robotic workers. “There are more sectors losing jobs than creating jobs. … If current trends continue, it could well be that a generation from now a quarter of middle-aged men will be out of work at any given moment,” he predicted.
After all, a worker laid off by a bot likely needs to be retrained to achieve a higher skill level than before to secure a new job. Increasing the educational level of an entire population, however, is a highly daunting and perhaps infeasible task. Training is costly and time-consuming, and moreover, as Harvard Business School assistant professor Gautam Mukunda said, “The bar for ‘this is how skilled you have to be to not be replaced by a computer’ keeps going up and up. It just might be reaching a point where it’s just not reasonable to have some people reach that bar.”
Besides demonstrating their value, new-age workers must also justify their cost, a task where automatons may again have the competitive advantage. Robots do not get tired and are immune from many human foibles—like greed, pride, and carelessness—that undermine businesses’ productivity. A 2014 report by the British international management consultancy Jomati Consultants LLP claims that robots will eventually match the work of “a dozen low-level associates” at law firms. Between ever more stringent eligibility requirements and a price war with machines, a large swathe of humanity may indeed become unemployable.
Is there any evidence that humans are losing out in a race against the machine? The data seems to undermine what Mukunda calls “the traditional economist’s answer [that] jobs will appear because when productivity goes up there are always more jobs.” In Rise of the Robots: Technology and the Threat of a Jobless Future, Martin Ford claims that the shift by which “tools that increase the productivity of workers [are] themselves turning into workers” is responsible for phenomena like the unprecedented zero net job creation of the first decade of the 21st century.
In an interview with the Harvard Business Review, MIT’s McAfee and Brynjolfsson similarly described how national wealth and productivity have grown while incomes and jobs have remained stagnant over the same period in many countries. “The fact that the middle class has been hollowing out in country after country [shows that] there seems to be a common underlying force that’s affecting all these countries,” Brynjolfsson said. “We think that force is technology.”
In the Shadow of the Iron Giant
With such a dizzying array of plausible arguments leading down different paths, it may seem like trying to predict the future as far as techno-unemployment is concerned is a futile exercise. But societal shocks are almost guaranteed to occur as a result of these technological changes. “In the short run we know with some certainty that some workers will be displaced by robots and by technology,” said Harvard Law School labor and industry professor Benjamin Sachs in a conversation with the HPR. “If robots become intelligent enough that … we do see a long-term displacement of human labor by technology, we need to rethink a lot of fundamental things about the way to structure work, the way we structure tasks, the way we structure the social contract.”
As Mukunda noted, the wave of robotization has already begun. Jeff Burnstein, the president of the Association for Advancing Automation, a trade association for promoting automated technology, told the HPR that even at present, “just about every industry is increasing its use of automation these days”—from cars and electronics to the aerospace and food processing businesses.
China Daily recently reported the start of construction for the “first zero-labor factory” in Dongguan that will scale down Everwin Precision Technology Ltd.’s human workforce by up to 90 percent, and Midea, a Chinese electronic appliance manufacture, plans to cut third of its 30,000 workers by 2018 in favor of automated systems. Even if Lawrence Summers is wrong that employing workers in automated industries is the central “economic challenge of the future,” he was certainly correct that the trend towards increasing automation is “inexorable and nearly universal” and that it will usher in upheavals that will be profound in numerous dimensions, maybe unimaginably so.
Update (11/25/15): This article has been updated from an earlier version.
Image Source: Mixabest/Wikimedia Commons