They don’t necessarily look like we thought they would, but the robots are already among us. They give us directions when we’re driving. By 2015, Governor Deval Patrick recently announced, they will collect all tolls in Massachusetts. Apple’s personal assistant, Siri, might as well be an early prototype for protocol droids like C-3PO of “Star Wars.” The drones America’s sending into battle could be precursors to the robot army that rises up against humankind.
As robots get smarter and more competent, will we benefit? Robot wars aside, economists, at least, have assumed the answer is yes. The less menial work humans need to do, the more we can focus on the creative and flexible work that humans excel at—jobs that involve talking, listening, selling, inventing, choosing, designing. Most textbook economic models economists learn in school assume that when new-fangled machines drive growth, everyone ultimately benefits.
But a number of economists have noticed that today’s technological changes are affecting workers differently than those models would predict. To the extent that the economy has rebounded, it has added fewer jobs than in past recoveries, as businesses take the opportunity to replace more people with machines.
Some experts, like Jeremy Rifkin at the Foundation on Economic Trends, have gone so far as to predict that we’ll soon be living in a workerless world. But certain people are going to lose their jobs before others—and economists are beginning to take a hard look at the workers for whom robots may be a more immediate menace. Edward Leamer, an economist who teaches at the UCLA Anderson School of Management, has argued that while information technology is creating opportunity for the most talented workers, it’s exacerbating income inequality by reducing the skills needed to perform most jobs. And now a disturbing new study by two researchers at Boston University and Columbia University is suggesting that fast-moving automation may come at the expense of another group of victims: our kids.
Part of the reason economists are rethinking the impact of technological change is that machines can do so much more than they ever could before—enough to replace human workers outright. Not everyone wants to use the word robots: As Leamer pointed out in an e-mail, it “suggests physical labor, as in robots are replacing humans in manufacturing.” But whether you call them robots, smart machines, or computers, they are replacing not only factory workers, but tax preparers, bank tellers, service workers, and other people in jobs for which humans once seemed uniquely qualified.
To Laurence Kotlikoff, a Boston University economist who’s interested in generational equity, the change is obvious. In the past, taxis might have replaced horse-drawn hansom cabs, but both required human drivers. The driverless cars being developing today will conduct themselves. Jeffrey Sachs, an economist who heads Columbia University’s Earth Institute, had made the same observation: “There are auto plants where you go where there are basically no workers inside,” he says. At one he visited a few years back, there was no mess, no grease—“just this humming sound of huge robots working.”
In the context of rising unemployment, these observations prompted Sachs and Kotlikoff to ask an uncomfortable question: “What if the Luddites are now getting it right?” they write. In a new working paper titled “Smart Machines and Long-Term Misery,” they pursue this question and reach a worrisome conclusion: Computer programs and robots may now be competing with workers in a way that leaves some groups—including supposedly tech-savvy young people—worse off in the long term.
It was clear to Sachs and Kotlikoff that something was messing with the wages of younger workers. While 60 years ago, the income of men between 45 and 54 beat out that of 25- to 34-year-old men by 4 percent, today the gap has grown to 41 percent. Young people might be comfortable with smartphones and fluent on the Internet, but that doesn’t necessarily mean they have the skills to land those jobs that robots haven’t made obsolete.
To explore this possibility, Sachs and Kotlikoff created “an admittedly highly stylized life-cycle model” that begins with the idea that machines and unskilled workers can easily substitute for each other. They assume young workers are unskilled workers: It takes time to accrue education and on-the-job training. With this framework (a variant on a standard economic model) in place, they run through a series of equations, the last of which indicates that, over the long term, when machines’ productivity goes up, young people have a harder time investing in their own skills and in the technology that’s boosting economic growth. In other words, when machines can easily replace lower-skilled workers, we move toward a world where every generation is actually worse off than the previous one—even as the economy grows.
The way this model plays out stacks up with other economists’ findings about the downsides of technology. “Tech progress can make the pie bigger yet still make a lot of people worse off. It’s the big paradox of our era,” says Erik Brynjolfsson, who heads MIT’s Center for Digital Business. Although he says he wouldn’t necessarily start with the assumption that young people are unskilled, he believes the paper is raising the right issues. In the center’s work, “we show some broadly similar things that technology can make people worse off,” Brynjolfsson says.
The muddier question is who those people are—and whether, as Sachs and Kotlikoff suggest, younger people are among the groups being harmed. “There is much to the Sachs and Kotlikoff paper, but I don’t view it mainly as an old vs. young effect,” says economist Tyler Cowen. It may not make sense to assume that young people are the unskilled ones: “The key question is who works well with computers and who is competing with computers,” Cowen says. Often, he points out, technological change favors the young—just look at Mark Zuckerberg—so it’s perhaps too simple to group young people as one unskilled mass.
Of course, Zuckerbergs are not the norm. “There are a lot of young people to whom this model actually applies. Probably more than half of young people,” drawn, for instance, from the two-thirds of 25- to 29-year-olds without bachelor degrees, says Sachs. “If our labor force was all young computer programmers, we probably wouldn’t be having this discussion, period.”
Sachs and Kotlikoff are economists, not revolutionaries; they aren’t Luddite enough to want to smash the machines. Instead, they suggest a mechanism for making sure benefits are more evenly shared—in this case, a redistributive tax. Other economists think that it’s still possible to get people the right education to do the jobs robots can’t.
It’s going to be a tough slog identifying those truly irreplaceable human roles, though. “I don’t know if we’re ever going to fall in love with machines,” Kotlikoff offers. But even that’s already happening: In 2009, a Japanese man married his virtual girlfriend.
Sarah Laskow is a freelance writer and editor in New York City. She edits Smithsonian’s SmartNews blog and has contributed to Salon, Good, The American Prospect, Bloomberg News, and other publications.