TCS Daily

Robot Nation?

By Glenn Harlan Reynolds - December 3, 2003 12:00 AM

A couple of columns ago, I wrote a piece called Kent Brockman on Unemployment, describing the impact of robots and automation on employment. In the comments section, someone posted a link to some things that the writer and founder of HowStuffWorks Marshall Brain has written. Brain thinks that we'll be losing jobs wholesale to robots in the very near future, long before things like nanotechnology have a chance to change the world:


In 2003 we are seeing the deployment of automated checkout lines in stores all across the U.S. This is the leading edge of the robotic revolution in retail. By 2015 we will start to see voice-recognizing robots helping customers in these stores, inventory-shelving robots putting the products out, cleaning robots sweeping the floors and the parking lots, cart robots bringing the shopping carts back into the store.... Robots will be moving in to make the completely automated retail store a reality in a 2020 time frame.


Brain is certainly right that automation is replacing all sorts of jobs. From pay-at-the-pump gas stations that make it unnecessary to have a human cash-taker, to ATMs, to automated grocery-store checkouts, low-skill jobs have already been replaced by machines to the tune of millions. You can argue about the magnitude -- the pay-at-the-pump machines came in after full-service gas stations were almost extinct, meaning that they didn't kill as many jobs as the pump-it-yourself ethos, which didn't require robots, did -- but not the existence of this phenomenon. All sorts of jobs that were being done by human beings not long ago are now being done by machines. (And sometimes it's because the people don't want to work -- after my carpet-cleaning service failed to show up for the third time, I bought a steam-cleaning machine and they became permanent victims of technological unemployment, at my house. And that's one case where I'll shed no tears for the jobless.)


But in a way, Brain's argument seems to raise as many questions as it answers. If we're losing jobs to robots now in large numbers, why isn't unemployment higher? After all, even through the recent recession we've had unemployment levels that would have been considered "full employment" twenty years ago, when ATMs were a novelty, and when people could still find full-service gas stations without looking very hard. People are obviously finding work somewhere, which means that someone must be creating jobs. And if so, doesn't that undercut Brain's prediction?


I don't know. It's easy to point out that people have been predicting the end of employment due to automation for a century or two now, and that so far they've been utterly wrong. But, of course, to say that they've been wrong so far isn't the same as saying that they'll be wrong forever. At a sufficiently high level of technology, most of us can be replaced by machines. (Of course, at a sufficiently high level of technology, the distinction between humans and machines becomes blurred from both directions, but that's another column). And yet, as a wise man once said, "people gotta have something to do." Unemployment matters.


In part the reasons are economic: Barring a radical change in the way society does business, working is the main way people make money. "Don't work, don't eat" may not really describe how our society operates -- in fact, it doesn't really describe it at all -- but work is the way that most people make a living.


But beyond economics, work is important to people in other ways. It's a major source of self-respect, of structure, and of constructive human interaction. Experience suggests that people who don't work tend to be less constructive: whether idle heirs and heiresses, members of the titled nobility, or welfare recipients on the dole, not working seems to make people, on average, less responsible and, in quite a few ways, less admirable people. A society of non-workers, even if everyone were rich and healthy, would probably not be a vital, healthy society.


That's why I don't think I agree with the redistributionist approach that Brain favors (he calls it an "economic security system" but it basically boils down to taxing corporations and giving everyone an annual income). In fact, I don't think I agree with his general approach to economics. But I think that important parts of his analysis are likely to be true: First, lots of low-skill jobs are disappearing forever. Second, that trend is likely to accelerate. Third, even if (as I suspect) the economy generates new jobs to replace the old ones, the new jobs may not be as low-skill, and they won't magically appear in synchrony with the disappearance of the jobs they replace. The upshot is that there are likely to be both economic and political repercussions from technological change, and the technological change that drives them is likely to occur at an accelerating pace. That will produce both short-term and long-term consequences.


In the short-term, we're likely to see a swing toward protectionism and perhaps even a growth of anti-immigrant sentiment; the former is already beginning to show up in the 2004 election dialogue, and the latter may still appear before things are over. In the longer term the consequences are likely to be more significant than Brain thinks, though it's hard to say exactly how.


And -- though this is really another column -- it's the accelerating pace, even more than the change itself, which makes predicting the future difficult. Brain's view of what will happen may turn out to be correct as far as it goes, but it involves so many other changes going on simultaneously that the overall picture is uncertain, even if you think you understand one factor such as the growth of automation. The writer Vernor Vinge coined the term "singularity" to describe the way in which these changes make the future unknowable. Technological unemployment, in other words, will be the least of our problems; vast, machine-created wealth will be the smallest of our opportunities. As I say, that's another column -- or fifty. Meaning that, for the moment, technological change is keeping at least a few of us employed.



TCS Daily Archives