Imagining the End of The Age of Labor

The tension between technology and work is at least as old as the economics profession itself. A question some people are asking now is: if computers run by artificial intelligence can do the job of humans, will work disappear someday?

Two economists are proposing a couple different scenarios in a new paper that is part science fiction and part mathematical models. In one scenario, lower-paid workers who are not highly valued by society – say, McDonald’s hamburger flippers – are more readily replaced by computers than a scientist searching for a cure for Alzheimer’s disease. This will drive down wages for a larger and larger segment of the lower-paid labor force.

In a second sci-fi scenario, machines run by artificial intelligence, or AI, will ultimately be able to do any worker’s job. In that world, work “would cease to play the central role that it currently plays in our society,” the researchers predict. A computer, they muse, could even stand in for a judge. Farfetched? An AI judge might be superior if it “make[s] more accurate and humane judgments than humans, leaving behind the noise, discrimination and biases that have plagued our justice system.”

There are a host of reasons to doubt work will disappear. The economists who reject this worst-case scenario argue that technology is not job-crushing but job-creating. Machines, they say, free up workers from one type of job but open up new opportunities. Only the nature of work changes. It does not disappear. After World War II, for example, new industrial technologies created jobs that lured farmers into the cities. Artificial intelligence shouldn’t be any different.

The authors of this new paper do concede that what they call the End of Labor is far in the future. Supercomputers capable of the most sophisticated AI are extraordinarily expensive. It seems more plausible that jobs involving simple, repetitive tasks will be the ones increasingly replaced by machines. This has already started happening as robots have moved onto factory floors.

But if workers of all types are eventually replaced by machines, how would they buy their groceries, cell phones, and shoes? Something would have to be done to replace their earnings and “avoid mass misery” and “political instability,” the researchers say. They propose a universal basic income.

That would be a big shift from major government programs to replace wages with Social Security and unemployment insurance, which are contingent on a person having worked. But some cities are experimenting with a universal income for residents, and Andrew Yang, as part of his platform as a Democratic presidential candidate in 2020, proposed that all Americans receive $1,000 monthly payments.

In an AI world, people can work if they enjoy it. But a universal income “neither requires recipients to work nor actively discourages them from working.”

That raises another problem. Who’s going to pay for a universal income? Fewer workers means less tax revenue for social programs. It’s also difficult to imagine Congress approving it.

So if government can’t or won’t provide the basic income that would be necessary, who would? The researchers propose instead taxing the profits of “the winners of technological change,” which I take to mean companies whose productivity would increase “by orders of magnitude.”

The solutions in the world these economists are imagining is just too futuristic to contemplate.

Squared Away writer Kim Blanton invites you to follow us on Twitter @SquaredAwayBC. To stay current on our blog, please join our free email list. You’ll receive just one email each week – with links to the two new posts for that week – when you sign up here.  This blog is supported by the Center for Retirement Research at Boston College.