A while back I went to a movie on a weekday afternoon. As usual I bought my ticket at the automatic kiosk. As I crossed the lobby I noticed that there was nobody else around, neither patron nor employee. Even the snack counter was unattended. Ordinarily there’s an usher waiting to tear your ticket at the far end of the lobby, but this checkpoint was also unmanned, so I shrugged and made my way down the hall to my theatre, where I took my place among the half-dozen other lonely mid-afternoon moviegoers. After a few minutes the lights dimmed, the doors automatically swung closed, and the ads started. “Holy crap,” I realized. “We’re in a giant robot.”
Presumably there were two or three people working at the theatre that afternoon, and I happened not to run into any of them. But it’s easy to imagine how the handful of remaining jobs for humans could be turned over to machines. The ticket-tearer could be replaced with a barcode scanner and a swinging door. The snack counter could be replaced with a bank of vending machines. The only person that’s probably non-replaceable, in the near term, is the poor schlub who cleans the bathrooms. But we’ll figure out how to put him out of a job eventually.
Andrew Coyne, writing in the National Post, isn’t worried for that out-of-work toilet scrubber:
Artificial intelligence, we are told, if it does not altogether enslave us, will at the very least make us economically obsolete. Already it has started to replicate tasks previously thought the preserve of the human mind, from legal drafting to investment advice. Call it the Robots Will Take Our Jobs theory.
Where the RWTOJ thesis falls down is not in the idea that there will be jobs lost, but in its unstated corollary, that there will be no jobs created in their place.
Coyne calls this the “lump of labour” fallacy, as in, there’s a lump of jobs that needs a-doin’, and if part of that lump is handed over to machines, the workers presently doin’ those jobs will be left with nothing to do. Here’s why it’s a fallacy:
[I]n fact there is no fixed amount of work to be done. There is no permanent list somewhere of all the goods and services consumers might want or the jobs that might be filled providing them. Consumers’ wants are limitless, as is human ingenuity: not only do we generally prefer more of what we already want, but entrepreneurs are constantly thinking up new wants we didn’t know we had. Much of today’s workforce is engaged in making goods and services that not only did not exist a century ago, but had not been imagined.
Likewise, Coyne tells us, future displaced workers will find new jobs providing goods and services we can’t even imagine yet. Maybe not the exact same workers – he acknowledges that the individuals thrown out of their jobs might be irreparably hurt. But overall, the number of new jobs created will more than compensate for those lost. It’s happened before and we can count on it happening again.
He might be right. By definition we can’t imagine the things we can’t yet imagine. But I don’t find Coyne’s reassurances all that convincing, probably because I don’t think the “lump of labour” fallacy accurately describes the pessimists’ main concern about the prospective obsolescence of human labour. Because, you see, it’s not labour that we think of as an inflexible lump. No, we’re much, much more pessimistic than that! Our fear is that we humans are the inflexible lump.
It probably helps to put this argument in the form of a diagram. Here we have a visual representation of every thing a human can do:
The diagram, obviously, is incomplete. The list of unique things contained within this circle is infinite – we could, for instance, say a human is capable of writing a novel, or we could get more specific and say a human is capable of writing Lucky Jim, or we could get more specific still and say a human is capable of pecking out Lucky Jim one-handed on a typewriter while simultaneously sipping whiskey and smoking a cigarette.
Some of those refinements add value – if you can write Harry Potter and the Goblet of Fire instead of Lucky Jim, you can earn a half-billion dollars instead of a few paltry million. Most of the refinements, like typing one-handed, or in your underwear, or in a pressurized dome at the bottom of the sea, add no value at all. While one can imagine an endless number of refinements to any conceivable human activity, nearly all of them are economically worthless.
Furthermore, while the number of unique things in the circle is infinite, the circle is not all-encompassing. Undoubtedly in the future people will be employed doing things we don’t currently have words to describe – amphibiating the sensoid bits on tri-dimensional metadroms, maybe. But whatever as-yet-inconceivable things we may someday do, they all lie within the set of things humans are capable of doing. That set is limited not by the human imagination, but by human biology. Our lifting capacity is constrained by having only two arms which can exert only a certain amount of force. Our ability to do mental calculations is limited by the number of figures we can retain in our short-term memories. As for amphibiating those sensoid bits, whatever the job entails we can be pretty sure of certain things it won’t entail, like breathing sulfur dioxide, or seeing frequencies outside the visible spectrum.
I foresee your objection. With the assistance of machines, you say, we can augment our lifting, calculating, breathing, and seeing capacities. That’s why we invented machines in the first place, starting with the simplest tools – like the sharp-edged stones our proto-human ancestors used to tear the flesh of animals. I suppose an apeman who subscribed to the “lump of labour” fallacy would have complained that sharp-edged stones were putting able-bodied apemen out of work – where once three or four workers would have used their nails and teeth to dismember a hartebeest, now the job could be done by one hairy fellow with a stone. Brighten up, Andrew Coyne would have told those unemployed apemen, now you can spend more time hunting and gathering. And he would have been right!
Let’s put, right next to a circle representing proto-human capabilities, a second circle representing all the things technology could do, circa 1 million BC. We’ll have to zoom in to see it:
I’ve made the little circle overlap the big circle, because right from the start, technology is impinging on the previously exclusive domain of proto-humans – the sharp-edged stones are displacing manual labour. But that’s all right, because there’s literally an infinite number of other things the displacees can do with the time freed up by the invention of the sharp-edged stone. Learn to make fire, for instance.
You can probably see where I’m going with this, so let’s fast forward:
You’ll notice that the circle of human capabilities hasn’t expanded since the arrival of Homo sapiens. Better nutrition and education may have made us a little stronger or smarter, but the child of one of our Neolithic forebears, transplanted to modern times and raised as one of us, would be well within current mental and physical norms – perfectly capable of manning the snack counter at the local multiplex.
It’s not greater innate intelligence that has enabled us to make machines of a sophistication that our ancestors couldn’t have dreamed of. It’s the fact that new technologies are built on top of existing technologies. You can’t invent a self-driving car unless you’ve already invented a car, which depends on the earlier invention of the internal combusion engine, and so on back to the discovery of fire.
Human capabilities, unfortunately, don’t accumulate in the same way.
So, with every passing year, the expanding circle of technological capabilities takes a bigger bite out of the static circle of human capabilities. Every year the number of jobs for which humans and only humans are qualified diminishes. There are still, to be sure, an infinite number of options within the narrowing crescent of skills exclusive to humans, but among those infinite options it becomes harder and harder to conceive of an as-yet-untapped skill that can be turned into an economically viable career. And the crescent keeps shrinking.
And since it takes a fairly high degree of imagination and inventiveness to think up an entirely new career that won’t immediately be at risk of being taken over by machines, the many, many humans who aren’t particularly imaginative or inventive – I count myself among that number – are left to compete with robots in the penumbra of activities where our capabilities overlap. How do humans compete? Since we can’t work more quickly, or more reliably, or put in longer hours than machines, we have to work more cheaply.
This is a great outcome for employers. You don’t have to actually bring in machines to replace your workforce of fragile, clumsy, illness-prone, emotionally unpredictable humans. The mere hint that they’re replaceable should be enough to subdue their uppitiness. And every year the machinery gets faster, cheaper, more reliable…
I imagine to a guy like Andrew Coyne, who probably hangs out with other high-IQ types capable of dreaming up new goods and services and turning them into profitable businesses, the limitations of the human lump are less apparent than they are to us middling-IQ proles out here in lumpenland. “Don’t worry,” he tells us, “I read in Fast Company that the tri-dimensional metadrom industry is going to need tens of thousands of trained amphibiators.”
“But what kind of education will that require?”
“Just a two-year diploma.”
“But I’m a thirty-nine year old long-haul trucker.”
“No problem, the government will help pay for your retraining.”
“But how long before they develop robot amphibiators?”
“A decade and a half, at least.”
“And what do I do then?”
“By then tetra-dimensional metadrom technology should be in full swing.”
“And they’ll still need experienced amphibiators?”
“Well, you might have to go back to school again…”
“You realize I became a long-haul trucker in the first place because I didn’t do well in school, right?”
“Or maybe you lost interest in school because you thought you could have a decent-paying career without a post-secondary diploma.”
“Yes, yes, that’s exactly the point I’ve been trying to make!”
“I literally can’t understand what you’re complaining about.”
“I can see that. Well, thanks for your pep talk anyway, Andrew Coyne.”