Automation Revisited

The cover of the Oct 23 New Yorker magazine shows a street person with his begging cup and his dog at his side.  The beggar looks despondent but the dog appears to be surprised by the robot dog on a leash held by a humanoid robot passing by. In fact, it’s all robots of various forms carrying various objects like coffee and smartphones --- not another human or real dog to be seen. One of the robots is dropping what, at first glance looks looks like several coins into the beggar’s cup but on closer scrutiny we realize they are not really coins as we know them, they are instead washers and flat gears --- obviously the standard currency in this hypothetical world.

Since the covers of the New Yorker are usually cartoons depicting an ironic situation, what’s the joke here? The irony appears to be that not only are we becoming more and more dependent on robots and more and more useless to one another, we are watching the steady march of automation gobbling up our jobs.   This is a fear that could be said to have started in earnest during the early 19th century when Joseph-Marie Jacquard invented an automated loom which resulted in riots by the local English weavers who saw (correctly) that this device from the devil would eventually put them out of their jobs. The leader of this insurrection was Ned Lud and so  the disparaging label “Luddite” became applied to anyone who rails against new technology whether or not it displaces workers.

In the early 21st century economists sought to allay this fear of automation causing jobs to be lost by claiming that, on the whole, automation creates more jobs than it displaces.

However, current research shows that this claim is an oversimplification. Job displacement is, as commonsense dictates, dependent on the job type. In essence, jobs requiring little education are much more likely to be replaced by robots. Of course, there will be exceptions. Compare two occupations: housekeeper vs radiologist --- which would you predict is most likely to be replaced by automation? Certainly the radiologist needs higher level skills than a maid, but a strong case can be made that it is the job of the maid that is more difficult to automate. For example how do you program into a housekeeping robot every situation which may arise and the actions to take when it does? For example, A human maid upon finding reading glasses next to the TV would put them back in their case. (“The Lights in the Tunnel: Automation, Accelerating Technology and the Economy of the Future” by Martin Ford.)

Many of the points made by Ford are echoed in the article in the 10/23 New Yorker article, “Dark Factory” (Can Humans adapt to a robot workforce?) by Sheelah Kolhatkar. She argues that since near full automation is inevitable, the best solution is to bite the bullet and pay people who have lost their jobs to automation so that they may continue to consume goods and services and keep the economy from collapsing.

But where will this “guaranteed wage” come from?  Kolhatkar suggests that the businesses which save money by replacing human workers with machines be taxed to pay the displaced workers. Understandably, there will be a massive pushback against this idea as it smacks of Socialism but she believes attitudes can and will change especially under massive economic pressure.

Thanks to my natural talent for procrastination, I see that the Nov 13 issue of the New Yorker has all of its Letters to the Editor devoted to this article. One of them,by Robert B. Price, ends his letter with the observation, “Sometime in the next twenty-five to two hundred years, A.I. will be capable of writing its own code, and will do it much faster and better than humans. We need to think very seriously about that, and take steps to avoid the scenario ... on the cover of the magazine” [ the parade of robots walking by the despondant  human beggar and his bewildered dog].

Fortuneately, there are computer scientists  who realize that when robots are capable of building smarter, better, and faster robots, generation after generation,  this will result in exponential growth of their capabilites. And while we humans can do the same thing by standing on the shoulders of scientists before us, it is at a much slower rate than the robots can progress. How long will they put up with us?

Something to think about indeed!