“Ultimately, it’s a virtuous cycle,” declares economics reporter Annie Lowrey in a Times Magazine piece on the job-displacing effects of automation technologies, “because it frees humans up to work on higher-value tasks.” The challenge today, she writes, a few paragraphs later, “is for humans to allow software, algorithms, robots and the like to propel them into higher-and-higher-value work.” The idea is an old one. Aristotle compared tools to slaves: both provide their masters with time for more refined activities. Thinkers as various as Marx, Keynes, and Oscar Wilde said similar things during the industrial revolution. It remains a common refrain today, as automatons and software take over more of the work people used to get paid to do. “We need to let robots take over,” wrote Kevin Kelly last year. “They will help us discover new jobs for ourselves, new tasks that expand who we are. They will let us focus on becoming more human than we were.”
There’s something deeply comforting about the notion that labor-saving technology inevitably pushes workers to higher pursuits. It salves our anxieties about job losses and wage declines — everything will work out fine, “ultimately” — while playing to our unbounded sense of self-importance. The ladder of human occupation goes forever upward; no matter how high our machines climb, there will always be another rung for workers to clamber to. But like many of the comforting things we tell ourselves, it’s no more than a half-truth. And when trotted out as a pat response to contemporary unemployment and underemployment problems, it becomes a dangerous fallacy. By promoting a reassuring fantasy about the future, it relieves us from grappling with the possibility that new, structural problems are opening up in the economy.
The problem with the endless-ladder myth begins with the fuzziness of its claims. What exactly is a “higher-value task”? Are we talking about value for the employer, or value for the employee? Are we measuring value in terms of productivity and profit, or in terms of worker skill and satisfaction? Not only are those two things different; they’re often in conflict. One way that a machine can improve labor productivity is by reducing the number of workers required to produce something. Another way is by reducing the skill requirements of the worker’s job and hence reducing the worker’s pay. As analyses of the employment impacts of industrial machinery show, the use of technology to automate a job tends at first to enhance the skills of a worker, making the job more challenging and interesting, but as the machine becomes more sophisticated, as more job skills are built into its workings, a deskilling trend takes hold. The highly skilled craftsman turns into a moderately skilled or unskilled machine operator. Even Adam Smith understood that machinery, in enhancing labor productivity, would often end up narrowing jobs, transforming skilled work into routine work. At worst, he wrote, the factory worker would become “as stupid and ignorant as it is possible for a human creature to become.”
That’s not the whole picture, of course. In evaluating the long-term effects of automation, we have to look beyond particular job categories. Even as automation reduces the skill requirements of an established occupation, it may contribute to the creation of large new categories of interesting and well-paid work. That’s what happened, as the endless-ladder mythologists will eagerly tell you, during the latter stages of the industrial revolution. The efficiencies of assembly lines and other mechanized forms of production pushed down the prices of all sorts of goods, which drove up demand for those goods, which led manufacturers to hire not only lots of blue-collar workers to operate and repair the machines but also lots of white-collar workers to manage the factories, design new products, market and sell the products, keep the books, and so forth. The resulting expansion of a consumption-minded, experience-seeking middle class ratcheted up demand for all sorts of other workers, from retail clerks to doctors and nurses to teachers to architects to pilots to journalists to government bureaucrats to etc. A virtuous cycle it most definitely was. What it wasn’t was the manifestation of some universal virtuous cycle, some inevitable dynamic in the economy. It was a virtuous cycle very much contingent on its time, and one of the most important of the contingencies was the limited capacity of industrial machinery to take over human work. Even a highly mechanized factory needed lots of people to tend the machines, and most professional and other white-collar jobs lay well beyond the reach of technology.
Times are different now. Machines are different, too. Robots and software programs are still a long way from taking over all human work, but they can take over a lot more of it than factory machines could. It seems pretty clear now that that’s one of the main reasons we’re seeing persistently depressed demand for workers in many sectors of the economy. What’s perhaps less well acknowledged is the spread of the deskilling phenomenon into so-called knowledge work. As computers become more capable of sensing the environment, performing analyses, and making judgments, they can be programmed to take over more white-collar skills. Professionals and office workers start to look more and more like computer operators, tenders of machines.
There will always be opportunities for individuals to design cool new products, make new scientific discoveries, create new works of art, and think new thoughts. But that says little about the prospects for the labor market in general. There’s no guarantee that the deployment of computers is going to open up vast new swathes of interesting, well-paid jobs the way the deployment of factory machines did. Recent experience suggests that computers may have very different consequences. What they seem to be particularly good at is concentrating wealth rather than spreading it, narrowing the work that people do rather than broadening it.
The language that the purveyors of the endless-ladder myth use is revealing. They attribute to technology a beneficent volition. It “frees us” for higher-value tasks and “propels us” into more fulfilling work and “helps us” to expand ourselves. We just need to “allow” the technology to aid us. Much is obscured by such verbs. Technology doesn’t free us or propel us or help us. Technology doesn’t give a rat’s ass about us. It couldn’t care less whether we have a great job, a crappy job, or no job at all. It’s people who have volition. And the people who design and deploy technologies of production are rarely motivated by a desire to create jobs or make jobs more interesting or expand human potential. They’re motivated, as Adam Smith also pointed out, by a desire to make money. Jobs have always been a byproduct of the market’s invisible hand, not its aim.
The biggest beneficiaries of the endless-ladder myth are those who have gained enormous wealth through the profit-concentrating effects of commercial computers. The myth helps them feel good about themselves. They, after all, are the ones who are setting in motion the virtuous cycle that, in the fullness of time, will bring us all to the nirvana of “higher-and-higher-value work.” It suits their business interests, too, by conflating those interests with society’s interests. Software and algorithms and robots will solve our problems, if we allow them to.
I’m not saying that it’s impossible that we’ll soon be blessed with all sorts of great new jobs. The world’s complicated; the economy’s complicated; no one knows what the future’s going to bring. I’m saying that we can’t take it as a given that that’s going to happen, and we certainly shouldn’t assume that machines have the best interests of workers at heart. Ultimately, it’s a virtuous cycle — except when it’s a vicious one.
Image of ladder-climbing robot: DARPA.
It seems to me that the target should not be one of middle-class jobs for all, but of wealthy incomes for all. Automation will eventually make the former difficult. Automation does not inherently conflict with the latter – if robots are to create most of the value, the question will be one of distributing the fruits of their labour.
I’m pretty sure Marx didn’t regard machines as the enablers in allowing people to become more human, and to focus on higher-level tasks — it’s the ownership of the means of production that enables that. If the capitalist owns the machine that the worker is paid to keep up with, the worker doesn’t benefit at all. But if the worker owns the machine, that’s the game changer. It seems to me like there’s some confusion about this point, and that people have the tendency to conflate the techno-salvationist vision of distributed DIY robotics and 3D printers and so forth with the idea that the technology is itself the thing that will transform the marketplace for labor. It’s not the technology — it’s the distributed production, and the artisanification (for lack of a better neologism) of technology.
Yes, the ownership of the machines, in Marx’s view, had to first shift to the workers, but once that happened I think Marx portrayed the machines as becoming the enabler of the workers’ multifarious self-actualizing leisure-time pursuits.
As someone who purveys automation for a living, the uncomfortable conclusion that I continue to reach is that a person’s living should probably not be tied so closely to their job. If I can build an automated solution that takes the place of 20 human workers, then that’s probably a good thing for the company. But it’s foolish to suggest that those people just need to be freed up to do something with higher value, because they might not have anything higher value to do (or the education/interest/skill/etc. to do it).
I don’t feel good about suggesting a team of 20 people be fired, and thus unable to feed their family. Continuing on in inefficiency is not good for the business either, so I’d rather have a solution that gets us what’s good for the employee and what’s good for the company. But you’re right to not believe the myth.
Nick, thank you for another great entry in your blog. You put into great English what I have often thought or felt, but could not put into words.
In the early 2000s, as I was in school and deciding what career to pursue, I came across articles about the off-shoring of jobs in Silicon Valley (and other places) at a break-neck pace. While not automation, the cause is the same: reduce costs, which, in turn, really only benefits those at the top. Personally, I considered it treason or sedition. I know that is strong, but all of that out-shoring of jobs, while supposedly necessary to keep the business running, harmed our country more than helped it.
Some say that we (in this country or globally) are aided or improved by the increase in automation (or off-shoring of jobs), which reduce manufacturing costs and increases the availability of cheap, “nice” gadgets like cell phones, tablets, laptops, computers, and even helpful medical devices. Many swear by there iPhones and iPads or Android devices; they could not imagine living without it. I agree; they are cool. Nevertheless, I maintain that those things, to use the vernacular, are “nice-to-haves” only. Sadly, our society on the whole, or at least it seems, has become more concerned with the progression of its tools than of its people. I have come to learn that no (new) technology or scientific discovery (though the latter is much more significant and useful than the former) will improve society. (Making life easier doesn’t translate to an improvement of individuals.) The change must be us, or in us, not our stuff.
Annie Lowrey and Kevin Kelly sound incredibly naive – don’t they remember those folks who told us decades ago that the wonders of the computer age would allow us to enjoy 20-30 hour work weeks and have more time for leisure and family? Uh-huh, sure. While I’m wondering what happened to my 20 hour work week I’ll go find John Henry and tell him to take his hammer and find more virtuous and higher-level work to engage in.