Martin Wolf’s recent article in the Financial Times on the threat (and promise) of automation starts out with an analogy (maybe more of a metaphor): are humans becoming as obsolete as the horse? Horses are no longer a major part of the economy, having been displaced by technology. Will humans suffer the same fate?
This is not the first time this metaphor was used. Back in 2015, Erik Brynjolfsson and Andrew McAfee did the same in a Foreign Affairs article provocatively entitled “Will Humans Go the Way of Horses?”
As I noted back then (and Wolf notes in his article), horses have not disappeared. They have transformed from a work animal to a recreational role. I would note also that other “obsolete” technologies have made the same transformation, such as wind powered ships aka sailing. And as I mussed in my earlier posting, I suspect that with the improvements in self-driving cars, driving a car will eventually become mostly a recreational activity.
Does this mean that we will approach Keynes’ vision of an age of leisure and abundance, as described in his essay “Economic Possibilities for our Grandchildren”? Probably not. Keynes was describing a point at which people are so wealthy that they need not work anymore. He argues that technical improvements and capital accumulation will solve the “economic problem,” which he defines as the struggle for subsistence.
Economic history has shown that such a point is unlikely. There is the distributional issue. Society in aggregate may be wealthy, but that does not mean everyone is. There is the production issue. Keynes never describes how wealth equates to output. And finally, there is the issue of human nature. We humans have a tendency to want more—well beyond the level of subsistence. Keynes recognized the importance of work in defining one’s self. Keynes’ notion of a 3-hour workday was to give people something to do—not because the work was needed. And he has mixed feelings about the importance of excess accumulation for our sense of self.
Coming back to horses, I would venture to say that many would be thrilled at the prospect of “going the way of the horse,” that is becoming creatures of leisure rather than factors of production. But humans are not horses. Humans have a complex bundle of needs and wants that goes well beyond subsistence. I do not mean this to ignore equine intelligence, including equine emotional intelligence. It is just to note that creating and operate a human socio-economic system that meets those human needs is very difficult.
Humans also have agency over their lives; horses do not. Thus, humans can make decisions about the course of their lives and intervene to implement those decisions. More than the economic argument, Keynes’ essay was really about that fundamental philosophical question of how to live our lives:
“Thus for the first time since his creation man will be faced with his real, permanent problem—how to use his freedom from pressing economic cares, how to occupy the leisure which science and compound interest will have won for him, to live wisely and agreeably and well.”
Keynes was optimistically guarded about this happening. On the one hand, based on what he saw as the advanced guard in this new era, the future was not bright: “To judge from the behavior and the achievements of the wealthy classes to-day [1932] in any quarter of the world, the outlook is very depressing!”
But he remained, somewhat surprisingly, hopeful:
“I feel sure that with a little more experience we shall use the new-found bounty of nature quite differently from the way in which the rich use it to-day, and will map out for ourselves a plan of life quite otherwise than theirs.”
That “little more experience” turned out to be a continuing Great Depression, the devastation of WWII, and the horror of the Holocaust.
Of course, the past century or so has seen a remarkable improvement in the material well-being of humankind. But we have not seen anything like Keynes’ “different plan” (the “counter-culture” notwithstanding). Nor are we likely to. As Brad DeLong points out in Slouching Towards Utopia,
“We do not use our wealth to overmaster our wants. Rather, our wants use our wealth to continues to overmaster us. And this hedonic treadmill is one powerful reason why, even when all went very well, we only slouched rather than galloped toward utopia.
Nevertheless, getting off the treadmill looks grim. Only a fool would wittingly or ignorantly slouch or gallop backwards to near-universal dire global poverty.”
I bring up this abbreviated discussion of the fate of humankind to show how the analogy/metaphor of horses and humans breaks down. Humans aren’t necessarily “going the way” of anything other than the way of humans. The fate of horses can tell us only very little about the fate of humankind. But a reorientation of the metaphor might.
Rather than think of horses and humans as parallel, let us go back to the relationship between the two: horse and rider. A good horseperson will tell you that the relationship is symbiotic. The rider is ultimately in control but the horse has a mind of its own as well. Anyone first-time greenhorn rider on a trail horse can attest to that. The relationship works when horse and rider act as one.
Compare that with the car and driver. Yes, the driver will tell you about working with car—which is mainly understanding and utilizing the car’s quirks and peculiarities. But no driver has had to deal with a car not wanting to go where the driver steers it (unlike a rider persuading a horse at full gallop to jump over a fence).
So maybe we should be thinking of automation and AI as more like a horse than a car. AI is a tool to be worked with like a horse. The horse and the rider each bring something to the relationship. And that relationship depends on the task. With this thought in mind, let us revisit the difference between labor-substituting automation and labor-augmenting innovation. Sometime you (the rider) are interested in a galloping with the winds as horse and rider work as one (labor-augmenting). Other times you are just interested (my riding experience) in getting out and seeing the scenery where a car can’t go without having to walk (labor-substituting).
I realize that this metaphor may be a little stretched. But the point remains: AI and automation are both labor-substituting and labor-augmenting. Which is which depends on both the nature of the task and the decision of the human agents involved. Our public policy needs to be able to deal with nuance rather than assume technology is one or the other. Or force a technology into one category or the other. I’m not sure our policy making has reach this level of sophistication (or if it even can).
Such may be the fate of humankind.