our robot overlords

Can Workplace Robots Get Along With the Humans They’re Replacing?

A BAXTER manufacturing robot on display at Innorobo International Robotics trade show on March 18, 2014 in Lyon, France.
Baxter, your new intern. Photo: Patrick Aventurier/Getty Images

Last week, while on vacation in Tokyo, I went to the Robot Restaurant, a much-hailed Japanese tourist attraction that features a 90-minute show with all kinds of utterly insane robotics. (Anthony Bourdain called it the “greatest show” he’s ever seen.) In the show, dozens of huge, remote-controlled robots shoot fire out of their mouths, zoom around the stage, dance alongside women in bikinis, mock-fight each other, and take part in musical numbers. It’s one of the strangest things I’ve ever experienced, and the message of the show — beyond sheer entertainment — is that robots aren’t scary, job-destroying things we should be afraid of. They’re our wacky, lovable friends.

That message is also bubbling up in Silicon Valley, it turns out, where tech companies have started trying to put a friendlier face on workplace automation.

This week, I went to the Interactive Collaborative Robotics Workshop, a San Jose confab put on by the Robotic Industries Association, a trade group that represents companies making robots for factories and other work environments. The overwhelming note struck by the conference, as James Temple noted in Re/code, was defensiveness. Robots aren’t stealing your jobs! They’re working alongside you, helping you do your job better! Nothing to worry about!

Most employees are like, ‘Oh, we’re being replaced by robots,’” said Esben Østergaard of Universal Robots. The truth, he added, is that “when companies buy robots, they’re creating more jobs.”

Let’s hear them out. According to collaborative robot companies, robots can work productively alongside humans on factory assembly lines, doing simple kinds of repetitive, injury-prone work while freeing humans up to do more creative and challenging tasks. (A “collaborative robot” isn’t just a PR euphemism; it’s a set of technological standards that applies to any robot that is capable of working in the same physical space as a human, without injuring the human or needing huge safety cages.) Companies like Universal Robots make industrial robots that can grab, twist, pick up, and perform other simple assembly motions, making them potentially extremely valuable to companies whose workers used to perform these tasks all day. But you still need the humans there, to oversee the robots and perform more involved tasks that require human judgment. It’s not zero-sum, in other words — robots make humans better at their jobs, and productivity zooms ahead, making everyone happier, healthier, and more profitable.

The problem with this rosy outlook, of course, is that it’s not the robot-makers who will decide how their products are used. It will be CEOs at the companies who buy the robots, who will inevitably try to cut their costs by substituting robot labor for human efforts. (If Acme Inc. puts 100 collaborative robots on its widget production line only to find out that those robots can make as many widgets as 1,000 humans with, say, only a 50-person supervision team, it’s not going to let the other 950 people sit around and continue to collect paychecks.)

So the odd thing about a conference like the ICRW, designed in part for companies whose workers fear losing their jobs, is that it puts robot-makers in the odd position of having to explain what their robots can’t do.

Right now, robots are toddlers,” said Zach Gomez, a field sales engineer with Rethink Robotics. Gomez was showing me Baxter, his company’s workplace robot, which has two arms and a tablet for a face. Baxter is relatively cheap — $25,000 for a basic model — and is safe enough to work next to humans. The company’s marketing materials say that Baxter is capable of “doing the monotonous tasks that free up your skilled human labor to be exactly that.” In other words, it can’t perform surgery, play a concerto, or conduct rigorous logical analysis. But it can be easily taught to package small goods, turn screws, and do other types of repetitive motions.

Despite Baxter’s limitations, Gomez told me that he’s gotten “a lot” of interest in Baxter from CEOs looking to replace their human workers, especially in cities with high minimum-wage laws like Seattle. “People say, ‘Sell me a Baxter because I have to reduce my labor costs.’” Gomez admits that while Baxter and other robots will almost certainly grow more sophisticated over time, their present capabilities are more limited than most people realize. “It’s not automatically Baxters in one door, people out the other,” he said.

Why is the robot industry so worried about seeming threatening? Maybe in part because in the face of future regulation and potential backlash, it’s strategically smart to reframe the debate as being more about augmenting human workforces than replacing them. And maybe also because, like everyone else in tech, robot-makers want to believe that they’re solving human problems rather than compounding them.

But if the intentions of robot-makers are to create human jobs, that message might be lost on executives who salivate at the prospect of adding nonhuman workers who don’t command salaries or health-care benefits, who can work 24-hour shifts, and whose output is precise and predictable. Foxconn, the notorious Chinese company that builds devices for companies like Apple and Microsoft, has plans to bring in a million robots in the coming years, and already, humans are feeling the toll. “There were about 20 to 30 people on the line before, but after they added the robots it went down to five people, who just pushed buttons and ran the machines,” one Foxconn employee told The Wall Street Journal. And it’s not just China — a 2013 Oxford study estimated that as many as 47 percent of American jobs may be at risk of being automated in the coming years.

In the face of numbers like those, it seems futile for the robotics industry to lean on “collaboration” as a solution to the anxieties many Americans feel about automation. Robots may indeed work effectively alongside humans in many jobs, and the spread of automation to industries beyond manufacturing will benefit some by increasing productivity and freeing up workers for more creative tasks. (I’ve argued that robot journalists are a good thing for human journalists, for example.) Tech optimists like Marc Andreessen have argued that advances in technology will create many more jobs than they destroy in the future.

But even Andreessen would admit that the robotics revolution will likely cause some short-term dislocation in the job market, and that many of the people who used to work at places like Foxconn will need to be retrained for other types of work.

We are already seeing the economic effects of robotics in the workplace — in discrete cases like Foxconn’s, and in broader trends, like falling median incomes despite vastly improved productivity. And we know that workplace automation will continue to take a toll on the labor market we have now, even if those losses are mitigated by more tech-enabled jobs being created in the future. The whiz-bang Robot Restaurant view of automation might be fun for a tourist stopover, and the “collaborative” view of the U.S. robot industry’s PR machine is nice to contemplate, but odds are that the realities of the robot age, at least in the short term, will be significantly harsher than either one.

Are Workplace Robots Your Friends?