Though I wouldn’t say there’s anything new John Lanchester’s review essay on automation and the future of the labor market in the London Review of Books, he does an elegant job of making the case for alarm. He cites a 2013 paper by the Oxford economists Carl Benedikt Frey and Michael Osborne, which estimates that 47 percent of present-day employment is “potentially automatable” over the next two decades. Even if Frey and Osborne are overestimating the speed of technological progress by quite a lot, even if machines outcompete humans in only half or a quarter of today’s jobs, we’ll either see rapid employment growth in jobs not susceptible to automation, like in-person services that by definition require a human touch, or we will see an explosion in the number of adults lacking marketable skills. Moreover, Frey and Osborne anticipate that trends in automation will primarily displace low-skill and low-wage jobs.
What Lanchester doesn’t address are the implications of this rather plausible scenario for immigration policies in the market democracies, an issue that we often discuss in this space. Briefly, it seems very likely that less-skilled immigrants will bear the brunt of automation. Of course, it’s possible that a ready supply of low-cost immigrant labor might delay automation for a while, as indeed it has in many industries, but the logic of Moore’s law suggests that it won’t delay automation for long. Isn’t it still true that an aging population will need home health aides and other paid caregivers? Yes, absolutely. But these jobs that remain resistant to automation will become more attractive over time as workers are displaced from other sectors.
Back in 2013, Neerav Kingsland, best known as a charter school advocate, wrote on the drivers of education reform in the years to come, and one of his stray observations was that if technology eliminates many middle-class jobs (a scenario different from Frey and Osborne’s, I should note), we can expect that “[t]he education sector will likely capture some of this talent surplus.” He envisions that we might need fewer educators, as technological tools drive productivity increases, yet the talent pool will be stronger, as people who might have once become, say, lawyers decide to give teaching a try. The result will be better teachers, which is of course a very good thing. It’s how we get there that’s unsettling.
Many of you will no doubt object that we’ve had technological panics before, and every time entirely new sectors have emerged that have generated significant job growth. I hope that this is indeed the case. Keep in mind, however, that stringent labor market regulation might stymie or even prevent the rise of new modes of employment. Political resistance to the 1099 economy, the transition from full-time paid employment to self-employment mediated by sharing-economy platforms like Uber and Instacart, is rising. Higher minimum wages will likely make some workers with low levels of literacy and numeracy less employable, even in the absence of automation. This technological transition is not happening in a regulatory or a political vacuum. Rather than thinking about how we might ease the way for the newly self-employed, or how we might make life easier for low-wage employers that invest in the human capital of their workers, there’s a tendency on some corners of the left to vilify those who are trying to make this new economy work. That strikes me as very foolish. This techno-panic is different not because it is possible to imagine that new sources of employment will emerge. It is different because the forces arrayed against business model innovation are more formidable than in the past.