Be afraid, be very afraid: The robots are coming and they will destroy our livelihoods!

This box is the seventeenth-century equivalent of a quadrotor drone. From

In 1642, philosopher and mathematician Blaise Pascal built a little brass box, with a handful of dials sticking out of it. Each dial would select a number, and by a clever mechanism, another dial would display the sum or the difference of all the numbers selected. Pascal had invented the world’s first calculator. This freaked people out. Math, at that time, was synonymous with reason, which was the main qulality separating human beings from the rest of the animal kingdom. Pascal had taught minerals, which were lower than any animal, to reason. Pascal’s contemporaries anxiously asked what place would be left for humans in a world where metal could think.

We never really got over this anxiety. We’re all familiar with the Hollywood flims in which robotic exterminators with the faces of Arnold Schwarzenneger or Hugo Weaving hunt down humans in various robot-dominated post-apocalyptic wastelands. Lately, our science fiction robots are a bit friendlier. Think TARS from Interstellar, or the protagonist of Pixar’s latest, Big Hero 6. Friendly or not, however, the robots are still threatening our place in the world. The friendliness of helper-robots almost makes them scarier than the Terminator, because it is that very quality which is increasingly threatening our ability to earn a living.

This is a topic that terrifies me, but also makes me a little bit hopeful. So last night I took a train to London to see an Intelligence Squared debate on the subject. The proposition was appropriately ominous: “Be afraid, be very afraid: The robots are coming and they will destroy our livelihoods.” What follows is an account of that debate.

The first speaker, Andrew Keen, is a bit of a contrairian about the internet. The moderator, journalist Zainab Badawi, called him “The Antichrist of Silicon Valley”, his latest books are titled “The Cult of the Amateur”, and “The Internet is Not the Answer”. I wonder what he would think of me, an amateur who uses the internet to write optimistic things about technology. Keen’s opening statement, in any case, set up the debate quite well. He introduced the Morovich paradox: Computers tend to be very bad at simple tasks, like welding and folding clothing, but quite good at complicated ones, like analysing market data. This means bad news for the various educated professionals in the audience, he argued. Robots are diagnosing illnesses, marking essays, and doing research for law firms. And, as I’ve pointed out, they will also be supplanting human taxi and delivery drivers in short order. That means that we face an unsavoury future with even worse economic inequality than we have today: There will be a small upper class of programmers and entrepreneurs, a huge underclass, and nothing in between.

The first speaker against the motion was Walter Isaacson, who has written a biography of Steve Jobs, among a few other things. He started off with the rather unfortunate statement that “the industrial revolution wasn’t that bad”. One wonders if he has ever seen Oliver Twist. His argument that the total number of textile workers increased during the industrial revolution might be accurate in a strictly factual sense, but does not account for the quality or geographic distribution of the new jobs. That’s why the luddites (with whom he naturally compared his opponents) were so concerned about technology: There might have been jobs in cotton mills, but they didn’t provide the same quality of life that the cottage textile industry had.

The second speaker against the motion was Pippa Malmgren; a former economic advisor to President Bush. I must confess that a bit of a pigeonhole started to form when I heard that, and that it rapidly began filling up with pigeons when she started rattling out cliched platitudes about how “most innovationis coming from small groups of a few people working out of a garage somewhere”, and how anyone willing to pull themself up by their bootstraps can be one of those people. Alternatively, she said, people can become welders, because apparently there will also be a shortage of skilled tradespeople. “In my experience”, she said, as she waved a quadrotor drone in the air, “robots create jobs”. As with Isaacson, there was little discussion about the details: How many jobs, exactly, will the robots create?

The best speaker by far was the second speaker in favour of the motion: economist George Magnus. His argument hinged on one fact: That while the automation of the industrial revolution was largely about replacing human muscle power, the coming revolution will replace human brainpower. The difference, he argued, is crucial. Now that we have robots that can mimic our mental capacities, we are unlikely to have very much left to do.

There was a bit of wrangling over the details of Magnus’ argument. The speakers against the motion made a few vague and unconvincing arguments about creativity and interpersonal skills. Isaacson made a good point that it’s a lot harder to teach a robot to do some tasks than it is to teach a human, but he neglected missed the fact that you only have to teach one robot, after which point its expertise can be replicated indefinitely. And robots are already painting and composing music. In the question and answer period, Malmgren made some clichéd arguments about the power and promise of technology, which elicited an applause from some of the audience. Magnus’ blunt but honest reply, which got a much bigger applause, was that “that’s a very romantic view, which I would applaud as well, but I just don’t think it’s true”.

In the end, I think, the question of whether the robots will take our jobs comes down to a few pretty basic questions:
1) What will there be left for humans to do?
2) How many such jobs will there be?
3) What wages and working conditions will the majority of these jobs offer?

The answers, I regret to say, seem to be as follows: Not much; not very many; and bad. The speakers against the motion had every opportunity to disabuse me of this view, but consistently failed to do so. Repeated assertions that automation has always created more jobs in the past simply miss the fact that history does not necessarily repeat itself, and we count on it to do so at our own peril. Even if things will work out in the long run, that won’t necessarily help the next few generations. As George Magnus said, quoting John Maynard Keynes, “In the long run, we’ll all be dead.” If Walter Isaacson and Pippa Malmgren were really making the best case that could be made for our economic security, then I’m afraid to say that our jobs are probably doomed.

I’m not letting Keen and Magnus entirely off the hook. Their solutions to the problem sucked. Magnus proposed various band-aid measures such as wider use of labour-intensive construction. Try selling that one to a property developer. Keen, when asked by an audience member whether this meant we should rethink the purpose of work, replied with a pithy response of “We need jobs to earn a living!” This, of course, entirely missed the point of the question, which was presumably that we should find a way to arrange our economy so that we don’t need jobs to earn a living. Both of the speakers supporting the motion saw the fear in automation, and skilfully dismantled the opposition’s arguments, but they failed to produce any hope. Maybe I’m just naive, but as I argued yesterday, automation could be good news if we can manage it right.

At the end of the debate, the speakers against the motion had 52% of the audience in agreement, but the speakers for it had won over more people to their side, so they were declared the winners. I think the real verdict appeared during the question period, when a mother or teacher who had brought five schoolboys to the debate asked the speakers what career they should be working towards. Isaacson and Malmgren said they should follow their dreams and get working on their own entrepreneurial projects in their backyard shed (or become welders), while the speakers in favour were more pessimistic: Magnus pointed out that tech firms today only hire very few people, while Keen argued forcefully that they should NOT follow the entrpreneurial dream, because that would be like staking their future financial security on a lottery ticket.

I’m inclined to agree with Keen. Because if our only hope for the future is to become successful entrepreneurs, then most of us are doomed. You can’t have an economy where everybody is a tech entrepreneur. Malmgren and Isaacson’s insistences otherwise are dangerous, because they offer an appealing palliative for those 5 boys and their classmates. Kids of that generation will soon have a fight on their hands over the structure of an increasingly automated economy. I hope for their sake that we are not taken in by vague and illusory promises that we can all be the next Steve Jobs. Nice though that might sound, ultimately it stops us from getting down to the tough business of adapting our society to a world in which robots are increasingly sophisticated, and there is less and less for humans to do.

Technological Momentum, Automation and the Cooperative Movement

I found this while looking for a picture for this post, and I just couldn’t help myself. Created by Zort. Go upvote this design so that the good folks at Queertee will put it on a T-shirt!

There’s a concept in my field of study called technological momentum. It’s a nice compromise between the overly simplistic positions that technology shapes society, or that society shapes technology. According to historian and theorist Thomas Hughes, technologies in their earliest stage of development are a bit like a ball that has just been nudged off a ledge: It’s only starting to speed up, and it’s comparatively easy to stop it or redirect it, as often happens. Technologies in a later stage of development, however, have begun to amass a collection of related technologies, infrastructures, policies, consumer habits, and cultural connections. They are like a ball at the bottom of a slope: It is moving much faster, and is thus much more difficult to stop or redirect, and can in fact knock into other things and move them around. What this means, in a nutshell, is that in the early history of an important technology, its development is influenced by society; while in the latter stages, society is influenced by the technology.

This has interesting implications for techno-utopians like myself: If we want new technologies to be a force for good in the world, then we have to get at them early, when we have a better chance of influencing the form and path that they will eventually take. It means that activist groups wanting to use technology as a political strategy need to get out ahead of it, adopt it, and consciously shape it into something that can have a positive social influence once it gathers more momentum. Imagine, for example, what might have happened if the Luddites, rather than breaking the new industrial machines, instead adapted them for their own purposes could use so that they, rather than wealthy capitalists, could benefit from the increased productivity. It might not have been possible, but had they pulled it off, the last two centuries of history might be a lot less miserable.

With that thought-experiment in mind, I’d like to turn your attention to the subject of cooperative grocery stores. Personally, I’m a big fan of them. I am an enthusiastic customer of Unicorn Grocery-a very middle-class organic grocery shop in my part of Manchester. It’s a bargain compared to most other organic grocers, and it’s generally a much more pleasant place to do one’s shopping than in a gigantic supermarket. Plus the produce is fantastic, and they’re the only place I’ve found where I can buy tempeh.

Don’t worry: I haven’t been captured by Big Grocery Cooperative, and this is not going to be a 1000-word glowing review of an organic grocer in South-West Manchester. But I had to introduce them because I’m interested in here is the conspicuous lack of technology that has become nearly ubiquitous in British grocery stores: automated checkout machines.

The phrase “automated checkout machine” might give you a bit of a shudder. The machines are still pretty flawed, to the point that the sentence, “unexpected item in the bagging area” has become a minor internet meme. So I don’t really miss the presence of checkout robots at Unicorn. The opportunity to talk to a human cashier is generally worth the slight loss of convenience. And of course, eliminating jobs through automation doesn’t exactly fulfil the mandate of a worker’s cooperative very well.

But sometimes when I’m waiting in the queue to pay for my groceries, I find myself wondering what Unicorn’s end-game is on this matter. The big supermarket chains, and indeed most businesses, are going to be pursuing more and more automation in the coming decades. Other businesses, such as fast food will probably follow. It’s not implausible that cashiers could be phased out altogether in some places. The rage-inducing problems with the checkout machines will eventually be ironed out, the machines will look more and more like a time and money saver, and less and less like a malfunctioning nuisance. Customers will adjust their habits accordingly, and come to expect to interact with a robot when they buy their groceries. They might even become uncomfortable with the idea of a human cashier. In other words: Automated checkouts are about to gain a whole lot more technological momentum.

This prediction might make you uncomfortable. It makes me uncomfortable. But while it’s not guaranteed (I’ve never claimed to be a technological oracle), it’s definitely plausible. If it does come to pass, what will become of the grocery workers’ coops that have thus far eschewed automation? They might be able to hold on to enough of their old customers to stay in business. I personally don’t expect to be wooed away from Unicorn by automated checkout machines. A growing mass of consumers, however, who are becoming more and more more accustomed to the automated approach, will start to become more hesitant to use cooperative grocers. This will make it hard for the cooperative movement to grow, which would be fine if the goal of the movement was to establish a few sustainably and cooperatively run grocers in local neighbourhoods. Cooperatives, however, are often presented as a model for a kinder and more just economy, and this can only be the case if the movement expands its appeal. Can we expect that to happen if these businesses intentionally keep themselves behind the technological curve?

There’s a pretty compelling counterargument to what I’ve just said, of course. What good will it be for the cooperative movement to gain the world but lose its soul? We can’t continue to see cooperatives as ethical businesses if they start firing their staff wholesale in order to embrace automation. My answer to this is that automation does not necessarily have to be a destroyer of livelihoods. That’s where technological momentum comes in. The people who have shaped automation technologies in the past have typically been aggressive capitalists, with no interest in the welfare of the workers that their technologies displace. But what if somebody else, with more of a conscience, took a more proactive role in deciding what automation looks like and how it is used? Could there be a different outcome?

Suppose, for example, grocery workers’ cooperatives introduced a few automated checkout machines into their stores, but did not lay off any employees. Instead, with some workers freed up from the need to man the till, they could use the same number of employees to do less work. That means easier work for the same pay. The time savings, meanwhile, could attract more customers into the store, resulting in more dividends for the workers as well as a little extra boost for the cooperative movement. I’m not an economist, and I’ve never worked in a cooperative, much less managed one, so there are probably serious flaws in the plan I’ve proposed above. Coming up with a real, socially just plan for automation has to be a job for people with more business and technological expertise than me. What I’m saying is that it’s not a discussion we’re having, and that needs to change. The best business minds of the cooperative movement should be thinking about this question. Maybe I’m wrong, and automation is inherently bad for workers. If that’s the case, I’d like to see a detailed argument as to why.

It’s actually quite a bizarre thing that our society is so skeptical of automation. If we lived in small communities based on local economies and cooperation, then any way to automate our daily tasks would be hailed as great news, heralding less work and more stuff. It is only because of the social and economic institution of employment, in particular the view that it is necessary as a precondition for survival, that makes us want the business of running our society to consume more person-hours, rather than less. If, as I have argued previously, the very idea of jobs is going obsolete, then we might be able, once again, to see automation as good news. We just have to find the right way to implement it. Automation isn’t going to go away. Too many powerful people are actively pursuing ways to eliminate more and more human workers for that to happen. If, however, we can find a way to make automation happen on our terms, then it could actually be good news.