In the recent presidential election, automation and robotics got a slight reprieve from the accusations that it has been the key driver in job losses in the United States. During the campaign, the conversation shifted, thanks largely to then-candidate Trump’s masterful scapegoating of Mexico and China, while calling out trade deals like NAFTA and the Trans-Pacific Partnership as clear and present threats to U.S. manufacturing.
Indeed, the administration continues to downplay automation as a factor in the U.S. economy, because that explanation runs against the political policies it hopes to enact under the guise of improving the conditions of America’s workforce. On Friday, Treasury Secretary Steve Mnuchin dismissed the prospects of artificial intelligence and automation eroding the workforce. In an interview with Axios, Mnuchin said:
“it’s not even on our radar screen…. 50-100 more years” away. “I’m not worried at all” about robots displacing humans in the near future, he said, adding: “In fact I’m optimistic.”
But even as some politicians look to divert attention from the issue, public focus returned to the evils of automation. The New York Times ran a story titled “The Long-Term Jobs Killer Is Not China. It’s Automation,” while the Associated Press explained “Why robots, not trade, are behind so many factory job losses.” You get the picture. Technology is killing manufacturing jobs.
And there’s truth in all of these reports. Robotics and automation have been linked to lost manufacturing jobs in the U.S., and even the most pro-technology industry analysts would have a hard time disputing it. But that simple fact raises some complicated questions.
Are we living in historically unprecedented times for job loss? Or is this part of a cycle that predates even the Industrial Revolution? Is it possible to retrain our workforce for these changes? Or will the gap between educated and non-educated workers only continue to grow?
This exceedingly complex issue has no simple solution, and any attempt by politicians to villainize or sensationalize matters will only serve to further its negative impact. Industry and government alike need to take a long, hard look at the effect of automation on industries as a means of maintaining the United States’ role in manufacturing and innovation, while stemming domestic job loss.
A number of representatives of pro-automation companies and advocacy groups I spoke with used words like “scaremongering” to describe a spate of recent reports that have raised alarm around the role of robotics in job loss. But when pressed, those same organizations will ultimately acknowledge that automation has been a driver of factory job loss in the U.S., at least in the short term.
It’s pretty simple arithmetic, and something we’ve witnessed time and again. A much-cited Ball State University study suggests that automation has already proven a major driver of job loss this millennium. The paper notes that the decade between 2000 to 2010 marked the U.S.’s largest decline in manufacturing jobs in its history.
Those numbers are supported in part, by statistics from the Bureau of Labor Statistics, which notes that manufacturing jobs in the U.S. increased between 1994 (the year NAFTA went into effect) and 2000. After that, however, things look decidedly less rosy, with a loss of five million jobs in the intervening years. In spite of this troubling stat, productivity actually rose, according to Ball State’s report.
That study only chalks up 13 percent of job loss to trade during that period, with automation constituting a major portion of that remaining job loss. “In 1998, the inflation-adjusted output per worker was much lower than it is today,” write the study’s authors, Michael J. Hicks and Srikant Devaraj. “This is due to a variety of factors, chief among them being the automation and information technology advances absorbed by these sectors over this time period.”
Colin Parris, the vice president of Software Research at GE, is refreshingly straightforward when speaking to TechCrunch about the topic. “Yes,” he says, matter-of-factly, “there will be job losses.”
It’s blunt, sure. But it’s refreshing coming from an executive at a company so heavily invested in automation. But Parris’ story, naturally, doesn’t end there. His long-term projections — and those of his peers in manufacturing — are actually a fair bit sunnier.
“The only way to fight [job losses],” Parris continues, “is to train the talent that we have. Because in the future, we have to embrace robotics. It allows us to reduce cost. If I reduce cost, I have more money that I can use for innovation. The more money I have, the more new products I can create. The more products I create, the more workforce I can hire.“
That’s a trend that certainly has historical precedent. Technology has had a major impact on the workforce dating back at least as far as the Industrial Revolution — when various tasks became more automated and the types of jobs available changed as a result. At the turn of the last century, 41 percent of U.S. jobs were in or around agriculture. A century later, the number had plummeted to 1.9 percent.
And while technology adoption and its attendant short-term job loss certainly transformed those historical economies, that evolution of work didn’t lead to mass unemployment as much as a transformation of the work being done. Changing times have traditionally closed doors and opened windows, as the old adage goes. As one representative of a major technology firm handily pointed out on background, the number of jobs that have been completely eliminated by automation boil down to one key position. It’s a sentiment backed up in a Boston University study released in late 2015.
“[T]echnology rarely automates major occupations completely,” writes the paper’s author, James E. Bessen. “Many occupations were eliminated for a variety of reasons. In many cases, demand for the occupational services declined (e.g. boardinghouse keepers); in some cases, demand declined because of technological obsolescence (e.g. telegraph operators). This, however, is not the same as automation. In only one case — elevator operators — can the decline and disappearance of an occupation be largely attributed to automation.”
So, if you work or live on a floor higher than three, maybe pour one out for the poor elevator operator, whose numbers seemed to have peaked at around 1950 when 97,000 were registered in the U.S. census and are virtually non-existent today.
If it’s assumed that job automation has and will continue (in the short term, at least) to result in some form of job loss, automation’s defenders point out that many of the jobs that will ultimately be displaced will be jobs that “no one really wants,” or at the very least, positions that employers have difficulty keeping staffed.
“It might take employees out of what we call the ‘three Ds,’ a dull, dirty or dangerous job,” says Bob Doyle of the Association for Advancing Automation. But “[it] puts them hopefully in a different position that creates more value to the company,” he added.
Parris also cites the “three Ds” — referring specifically to flare stacks used to burn off the flammable gas from drilling operations in the Bering Sea. “These flare stacks are exposed to the elements because they’re out in the ocean, and you have to have people climb these things and look to see if there’s rust and corrosion,” Parris explains. “Who wants to do that? They’re dull, dirty and dangerous. It’s a huge problem.”
On a recent visit to the Massachusetts offices of industrial robotic gripper maker Soft Robotics last month, I put the question of replacing human labor with robots to CEO Carl Vause.
“There’s a labor shortage in packaging, so we’re working with customers in companies where they don’t have enough labor to staff their factories,” he answered, he says. “A lot of these companies have temporary labor constantly coming and going and worrying about training and quality and all of those challenges. So, what we’re looking to automate is packing dough in a 40-degree warehouse. It’s not a good job, it has a high turnover.”
However, MIT economist David Autor takes issue with the notion that automation is simply replacing jobs that no one really wanted in the first place. While most of the roles displaced by technology were not particularly glamorous, they were still staffed by humans requiring steady employment to make ends meet.
“I don’t think that automation will mean the end of work, but I do think the distribution consequences have been quite significant, and I do worry about those,” Autor says. “I think the set of good jobs available to low educated or non-college educated workers without specialized skills is definitely contracting.”
Autor, while not as apocalyptic in his vision as some of his peers, does consider growing professional dependence on machines a major contributing factor to the growing income gap between rich and poor. And it’s a vicious circle, as growing unemployment makes it difficult or impossible for workers to enroll in the education required to break out of poverty.
“I think automation has had an absolutely enormous effect on the structure of work, the returns to skills and what people do with their time,” says Autor. “I don’t think it’s created mass unemployment so far, and I don’t think it will. But it has dramatically reduced the set of opportunities available to uneducated workers. And I think that’s having very significant social and even political consequences.”
Short versus long term
The conventional wisdom among economists holds that, while there may be some short-term job loss at the hands of automation, the vacuum will ultimately fill in over time. The notion echoes the historical precedent of things like the agriculture example. And there is probably something to the notion in some cases that increased productivity and margins provide companies with funds that allow them to expand their size and employment ranks.
“There are a lot of middle-income companies that, rather than getting rid of people when they get robots, they realize people are of value, especially if they’ve been working for them for a long time,” says Doyle. “There’s always some concern when people bring in a robot. But I think once they realize that it’s there as a tool to help their companies, they really embrace it. Some of these companies… they even name the robot.”
It’s a rosy picture, but there’s a clear caveat. As the market shifts, even the most labor-friendly companies won’t be able to find new roles for the entirety of their workforce. Doyle describes different new roles that are being opened up in factories alongside robots, but the majority seem to require, if not an engineering degree, some sort of technical skill.
“The labor market for college-educated workers is very, very strong,” says Autor. “And those people continue to get paid better. The set available to people who just have a college education or less has dramatically contracted as a function as well as trade, but automation has been the bigger fact. It’s an important part of the growth of inequality we’ve seen over the past several years, the decline of earnings and fortunes of people without a college degree.”
It’s a deepening divide that speaks to the ever-growing need for STEM-based education that goes well beyond higher education. Common wisdom holds that the addition of productivity and the lowering of costs will open the possibility of more employment. And indeed, as factories become automated and more robots are introduced into the process, new, unforeseen categories will open up in the workforce.
But unlike some of the more entry-level manufacturing positions, companies do often have a problem filling roles that require a college education, or at the very least, specialization. As the nature of manufacturing jobs shift, so too will the divide.
While as Doyle suggests, it’s in the best interest of companies to reward loyalty among employees, how many corporations will ultimately take on the financial burden of educating a changing workforce?
And even then, how much of a dent can such education really make? Larry Summers is skeptical. The National Economic Council Director turned Harvard professor who penned the bluntly titled piece “Robots are hurting middle class workers” tells TechCrunch, “I’m sure more and better STEM education would be good but doubt it is the panacea some suppose. If STEM were in genetically short supply two-thirds of engineers would not be working outside engineering and there would be more acute wage pressure across the board.”
Here, again, historical precedents can provide a little peace of mind. In his 2015 essay, “Why Are There Still So Many Jobs? The History and Future of Workplace Automation,” Autor cites an awfully familiar sounding 1961 Time Magazine story titled “The Automation Jobless,” in which the author writes:
“In the past, new industries hired far more people than those they put out of business. But this is not true of many of today’s new industries. . . . Today’s new industries have comparatively few jobs for the unskilled or semiskilled, just the class of workers whose jobs are being eliminated by automation.”
Perhaps the piece jumped the gun by half a century. Or maybe American ingenuity will simply find a new way to self-correct. There’s hopefulness to be found in amongst all the doom and gloom of eroding manufacturing numbers.
“There’s a lot of scaremongering stories about jobs that will disappear in 2020,” Neil Kinson, Chief of Staff at Redwood Software, tells TechCrunch, “but the reality is actually that there will be jobs in 2020 that my children will move into that we don’t even know exist today, as is paralleled by any of the great technological advances, whether it was automating agriculture, looms, the Industrial Revolution, whatever it is. What we’re seeing is the ability to develop new services that drive growth and employment.”
A newly released study conducted by the Centre for Economic and Business Research (CEBR), in conjunction with Kinson’s company, notes that “A review of recent related economic studies suggests that there is a positive relationship between robotics automation and economic development.”
And there does seem to be a potential silver lining in the form of nearshoring — bringing manufacturing closer to the userbase of its output.
“Natural disasters such as the Tsunami in Japan can knock out a supply chain that is insufficiently diversified, making redundancy in the system and the capability to manufacture in a number of places more attractive,” Columbia Business School associate professor Rita Gunther McGrath wrote in a piece for The Wall Street Journal. “And as we see advances in automation and digitization, personnel cost as a fraction of total value created can be decreased, again making the economics of offshoring less compelling.”
Automation and robotics could play a major role in making that ideal a reality. Technology has the potential to help drive down the cost of production and could ultimately even help domestic manufacturing cost competitive.
An AlixPartners study from January notes that “A decade ago, the labor-cost advantages of offshoring were clear. Today, determining the strategic location of manufacturing assets is a complex, high-stakes process.” Automation is, undoubtedly, one of the key differentiators, “Two-thirds of respondents say they plan to invest a significant portion of their future capital in robotics and automation technologies[…] The case for those investments is strong: robots and automation technologies enable companies to reduce labor costs, enhance quality control, and improve throughput.”
Of course, even if it does help bring manufacturing back to the U.S., nearshoring certainly won’t mark a catchall solution to the recent loss of domestic manufacturing.
“I suspect we will see re-shoring of manufacturing as U.S. firms replace workers with technology and so labor costs diminish in relevance,” says Summers, “but this won’t help raise employment very much.”
If automation does result in a return of manufacturing to the U.S., even by the most optimistic accounts it will replace jobs once reserved for humans. But many involved in the industry do point to a shift in the way industry views robotics.
Collaborative robots are being viewed as a major growth potential for the industry, leaving the machines to handle the three Ds, while a human coworker takes on tasks that are still better managed by people. “Before with robotics, we would be trying to create autonomous robots that do things themselves,” says Parris. Now a lot of it is remote control … That’s new jobs that are being created.”
Collaborative robots tend to be more affordable and safer than their fully automated counterparts. It’s a concept that Amazon has deployed for some time now… at least since its acquisition of Kiva Systems, whose robots move shipping pallets in the company’s massive fulfillment centers. “We like to think of it as a symphony of software, machine learning, computer algorithms, and people,” a spokeswoman for the company told MIT in 2015. “And the people are such an important component; the technology wouldn’t mean anything if you didn’t have great employees that help interact and engage with it.”
With or without job loss, companies are going to continue to embrace robotics and automation in a major way. The bottom line for most corporations is that the bad press that comes with the elimination of what are generally regarded as low-skilled positions is offset by the savings, productivity and safety technology can provide.
But a future in which automation is ubiquitous in industries doesn’t have to mean an apocalyptic elimination of jobs as we know it. IBM, predictably, has seen a silver lining in automation’s influence on an evolving job market. The company’s CEO Ginni Rometty recently began discussing the concept of “new collar jobs,” in reference to positions augmented by things like automation and artificial intelligence.
In a letter to then-President-elect Trump, Rometty wrote, “The country should focus on infrastructure investments that incorporate Internet of things technology and artificial intelligence to improve performance.”
In a conversation with TechCrunch this week, IBM exec Guruduth Banavar echoes the sentiment, referring to the newly created category as “differently skilled” jobs, rather than lumping them into the standard high- or low-skilled categories.
“If you were in the same room with me and I had a whiteboard, I’d draw a graph,” he explains. “If you think of the distribution of expertise as a bell curve. If the high-end of expertise is on the right side and the low-end is on the left, there’s a whole lot of medium-level expertise in the middle. That’s going to shift to the right. What we consider today to be high-level expertise jobs will be taken by people who use machines in the future to fulfill some of the expertise needed.”
It’s a hopeful sentiment — that technology might help create new job categories, even as it displaces others. Banavar, naturally, cites the role IBM conceives of its Watson technology playing in a diverse range of fields, from medical to tax prep, augmenting the knowledge base of its users in the process.
And indeed, it reflects a historical precedent that we’ve seen played out time and again over practically every industry. Decades of science-fiction have taught us to be wary of robotics, and changing industries have only made us more keenly in-tune with those fears. But just because they don’t represent end-times, doesn’t mean they’re not real.
“It’s not just factory jobs,” says Autor. “It’s production, its clerical positions. I think there will be fewer sales workers going forward. I think retail positions will likely contract because people buy stuff online and it’s less labor intensive. I think the set of good jobs available to low educated or non-college educated workers without specialized skills is definitely contracting.”
When it comes to politics and economics, it’s easy to engage in scapegoating. In the case of manufacturing, immigration, trade and automation, all have been on the receiving end of plenty of scrutiny over the years. And automation can certainly shoulder its share of the blame for real-world job elimination, which, as evidenced by the above BLS numbers amount to a not-insignificant five million lost manufacturing jobs since 2000.
But technology has also been a major driver in helping keep companies competitive, so to shy away from it would surely only result in even greater domestic job loss. In order to move forward, we need to embrace technology both as a means of production and a method for producing new roles.
But pulling off such a coup is going to require some massive investments in education, both on the part of the corporations looking to move valued employees into new roles and an education system preparing workers for the real world. Failure to do so will only accelerate the growing rift between so-called low- and high-skilled workers, and the whole of our economy — and future — will suffer as a result.