Kay Hymowitz has written a piece in City Journal that is well worth your time. Some extracts:
People have feared artificial intelligence since Mary Shelley introduced the world to Dr. Frankenstein’s hideous creature. The Luddites, who battled against the automated loom in the early nineteenth century, are now regarded as so wrongheaded that they have an economic error named after them. The Luddite fallacy refers to the fact that in the long run, disruptive technologies create more jobs—not to mention reduce drudgery, save lives, expand leisure, and enrich us all.
Well, the Luddites were not quite as wrong as we like to think. Yes, in the long term the disruptive technologies of the industrial revolution created jobs and increased real wages, but it took over half a century — the so-called Engels (yes, that Engels) pause — for the standard of living of most British industrial workers to start increasing, too late for those who had taken the hammer to the looms.
The pace of AI discoveries and implementation is accelerating. Robots are now doing things that seemed like science fiction just a short time ago. Was anyone talking about a retail-sector meltdown, driven in good measure by AI-facilitated e-commerce, last year? Second, fasten your seatbelts. Whether you call it “the second machine age”—as MIT professors Erik Brynjolfsson and Andrew McAfee do, in a 2014 book by that name—or the fourth industrial revolution, this will be big. Most Silicon Valley honchos, scientists, and economists think that this time is different. Exactly how many jobs will be lost, which kinds of jobs and when, and what to do to prepare for these losses may be matters of dispute. No longer questioned is that a massive disruption in the way we earn a living is coming and that it will transform communities, education—and perhaps even our notion of an America defined by industriousness and upward mobility.
And if there’s an Engels Pause 2.0?
The piece is not all pessimism, far from it: Hymowitz gives plenty of examples of what she terms “technology’s counterintuitive influence on the labor market”:
James Bessen, an economist at Boston University School of Law and what you might call a lukewarm optimist, has studied the ATM’s impact on bank jobs. On first thought, it seems obvious: ATMs would mean fewer jobs for bank tellers, right? No, Bessen says. The number of bank tellers rose in the three decades after the ATM was introduced. The reason: along with deregulation, the ATM helped banks open more branches. Tellers were “up-skilled” to do more customer service and sales.
But what are those bank tellers paid?
Here’s the Washington Post from 2013:
Almost a third of the country’s half-million bank tellers rely on some form of public assistance to get by, according to a report due out Wednesday.
Among the vulnerable jobs listed by Hymowitz are truck drivers, train drivers, cabbies, fast food workers, and those with jobs in retail:
According to the New York Times, 89,000 workers in general-merchandise stores have lost their jobs in the past five months alone—more jobs than in the entire coal industry.
And those retail workers aren’t going to be able to take refuge in Amazon’s warehouses for long:
The many companies specializing in warehouse robots are already making ones that can lift cases off shelves. The machines need less space to get around than humans, so warehouses will no longer need to be stadium-size. The robots move faster than humans, too, and they can work 24/7. In 2012, Amazon purchased Kiva, a Boston-based robotics firm, for $775 million. Last December, the Seattle Times reported that Amazon had expanded its army of robots to 45,000—a 50 percent increase from the year before. The writing is on the warehouse wall.
The list of those at risk goes on:
Coal miners, steel, oil and gas, and construction workers… accountants, insurance-claims adjusters, travel agents, bookkeepers, translators….pharmacists, and even radiologists, journalists, lawyers, and Wall Street traders.
And let’s not forget paralegals.
Engels Pause 2.0 is looking more and more likely. And it may last even longer than last time round, taking a bite far higher up the food chain as it does so, something that is very likely to increase the political dangers of what lies ahead.
Moreover, most of the tech community and economists believe that income and wealth disparity will be as prominent a feature of an AI economy as it has been in a digital one. Will the prosperity they predict from the AI revolution be even more concentrated among a select few? What will be the social and political reaction to growing and seemingly intransigent inequality? Will the pressure mount for more redistribution? If, as seems likely, the answer to that last question is yes, conservative policymakers in particular will find themselves with few appealing responses.
And on that cheery note….
Update (via the Wall Street Journal this morning)
Nearly 16 million people, or 11% of nonfarm U.S. jobs, are in the retail industry, mostly as cashiers or salespeople. The industry eclipsed the shrinking manufacturing sector as the biggest employer 15 years ago…
“The decline of retail jobs, should it occur on a large scale—as seems likely long-term—will make the labor market even less hospitable for a group of workers who already face limited opportunities for stable, well-paid employment,” said David Autor, an economist at the Massachusetts Institute of Technology.