Ever since the Industrial Revolution, people have worried about machines taking jobs away from humans. Fears about employment track with every new technology, from the first automated weaving looms in the 1820s to the dawn of the Internet 20 years ago. But most people would agree that advances today create a new form of threat to work. Robots and other artificially intelligent beings are a different order of machine—less appliances that make things easier, more a wrench in the very wheel of life.
Various studies of the impact of robots paint a depressing picture of the future. For instance, one paper from researchers at Oxford University predicts that 47% of U.S. jobs are at "high risk" of computerization over the next two decades. All manner of positions could fall by the wayside, including jobs in transport and logistics, construction, mining, food preparation, and the police force. Even roles you might think of as "high value"—like doctors and lawyers—could be undermined that study found.
What will these people do? Perhaps, if they're inventive and hardworking, they'll find other productive activities. Perhaps we'll just do less and get paid less. In the last 15 years, we've already seen a trend towards less employment as automation hollowed out jobs in several industries (particularly the lower-skill white-collar type). In the future, this will probably continue and begin to affect all of us—even those lucky enough to work.
These trends are reflected again in a new paper from economists at Boston University and Columbia University. The study finds that "smart machines" will cause a "long-run decline in labor share of income," cause more technology boom-busts, and a "growing dependency of current output on past software investment." In other words, there will be less need for fundamentally new code to run computers and therefore less employment even for coders.
The researchers—Seth Benzell, Laurence Kotlikoff and Guillermo LaGarda from Boston University and Jeffrey Sachs from Columbia—argue that robots will increase wealth inequality further and that we'll need big redistributive policies to even up the lives of people with work and people without. "Our simple model illustrates the range of things that smart machines can do for us and to us. Its central message is disturbing," they say. "Absent appropriate fiscal policy that redistributes from winners to losers, smart machines can mean long-term misery for all."
The working paper, published by the National Bureau of Economic Research, is based on a computer model that simulates the real economy and features two types of workers—high-tech and low-tech—each of whom consume goods and services for two different periods. The high-tech workers produce code, which they license for immediate use and by selling rights to future use. The low-tech workers are artists, musicians, priests, and psychologists who put no capital into what they do; their only input is their labor.
The authors argue that platform-type computer programs will lead to less work for the coders over time. "Current investments in designing self-driving cars and other smart machines are monumental projects," Benzell, a PhD candidate in economics, says in an email. "Once a smart machine is designed that can drive a car, this is basically a 'solved problem.' Maybe it needs to be updated or tweaked a bit periodically. But the point is that the 'output' of car-piloting will be produced by past code rather than the labor of new workers."
At the same time, the very thing that makes coding attractive as a job—that you can earn lots of money doing it—makes it vulnerable over time. "Programmers get big wages to do this new, highly productive thing. But as programs start to accumulate, the next most useful thing to program becomes less useful. Because prices are set at the margin, this means wages for programmers, and in some cases, everyone, decrease," says Benzell.
The economic problem is that the people who gain from "higher code retention"—that is companies that own the smart-car platform and have to make just the occasional tweaks—stand to reap lots of money, while everyone else is "immiserated." That's why the paper calls for policies like higher capital income taxes (with proceeds going to a national investment fund), a revenue income tax (with proceeds going to young savers), or—more conservatively—less welfare for old people who've managed to build up wealth, and more for younger people who haven't and now have to compete with robots.
The paper makes a profound point. Wherever you stand politically, it's clear that the basic contract of working hard and getting by is breaking down. For technological and other reasons, employment isn't what it used to be. In the future, if we don't want legions of unemployed, we're going to need new policies to compensate.