Five Questions for David Autor
During the presidential campaign, immigration and free trade were largely blamed for stealing American jobs. But more recently, another boogeyman has been gaining attention: the dismayingly capable robot.
Automation, amped up by artificial intelligence and its potential impact on future employment has sparked research, debate, and disagreement among a wide range of stakeholders. Consulting firm PwC is the latest to jump into the prediction game, arguing in March that machines could take nearly four out of 10 jobs in the U.S. by 2030. By contrast, in the same month, President Trump’s treasury secretary Steve Mnuchin said that he doesn’t expect robotic machines to replace humans at work until a good 50 to 100 years from now.
To sort through some of the debate, Undark contacted MIT labor economist David Autor, who has spent the last decade studying the impact of technology and trade on unemployment, and whose work highlights a key paradox of modern technology — one that runs counter to common wisdom on the matter: While machines are increasingly involved in our work, we don’t seem to be running out of jobs.
We asked Autor what he makes of the recent predictions and prescriptions. Responses have been edited for length and clarity.
UNDARK — As someone who studies the impact of automation on jobs, what do you make of the predictions that have come out in the last few years about how computers and robots will affect jobs in the future? In particular, what do you think of research from University of Oxford economists Carl Benedikt Frey and Michael Osborne arguing that 47 percent of jobs in the U.S. are “susceptible to computerization” in the next 10 to 20 years.
David Autor — The predictions are rather sensationalist.
It doesn’t really make sense to ask whether automation will affect jobs. Yes, 100 percent of jobs will be affected. Jobs change all of the time. The content of jobs will change. But it’s not as if there’s a fixed amount of work to do. The net number of jobs is rising.
Job tasks are changing. In many cases that automation is complementary to the tasks that people do. For example, doctors’ work is becoming more automated, but that doesn’t reduce the need for their expertise. (For instance, testing gets automated, but that generates data that doctors need to interpret.) So, the impact of automation is much harder to predict than any of us have a handle on.
UD — So, it’s safe to say some workers are more vulnerable than others?
DA — Yes, that’s true. For the middle skills and trades, we’ll have fewer factory or clerical jobs, and but will likely need more medical technicians, skilled tradespeople and skilled repair people.
Our weak K-12 system makes this [situation] harder. Adults aren’t that flexible, and they’re harder to retrain. We’ve let the working situation for a good portion of Americans deteriorate and now we’re paying the price for that.
Education is critical. We should invest in our skills in ways to keep us relevant. A strong back and good character are no longer enough. It’s very likely that the jobs of the future will require expertise, judgment, and creativity.
Without knowing what the world will look like and where we should specialize, we should focus on fundamental skills in K-12. Those are reading, math, analytical reasoning, communication, and teamwork.
UD — In the recent U.S. presidential election, the candidates and voters spoke about the loss of manufacturing jobs due to globalization and immigration. For the most part, no one blamed robots. Why do you think that is?
DA — Automation’s effects have been much more gradual. Trade shocks [such as the effect of low cost labor from an emerging market] have been more obvious. Also, Apple and Google are American companies and people feel like they get some benefit from that. Also, people feel like they don’t have a vote when it comes to technology.
UD — That’s an interesting point. It seems like we see many more technology policy prescriptions coming from tech people, rather than policymakers. For example, Microsoft co-founder Bill Gates recently said we should tax robots to slow their adoption and earn revenues to offset their impact. What do you think of that idea?
DA — Treating robots like a negative externality is counterproductive. If we don’t develop the robots, we’ll soon be buying them from Germany and China. A better option might be to create a sovereign wealth fund on patents… I’m not saying the government should take sovereign control of Google, but a lot of patents come from publicly-funded research. The government invests a lot in technology development through DARPA (the Defense Advanced Research Projects Agency). You could say the government should be the residual claimant on some of the intellectual property created from that research. Economist Michael Kremer talked about this in the context of pharma patents. The government could buy a stake in some of the intellectual property created from that research and use the proceeds for public investment.
UD — Some futurists talk about a time when we won’t have to work because so much will be automated. Some people will get insanely wealthy and others will need a universal basic income to survive. What do you make of that? Does it seem realistic?
DA — We already work less. We used to work six or seven days a week. We could retire earlier but we choose to consume more. As we get wealthier, we’ll have even more consumption… So how do we assure that more people benefit?
I’m not sure universal basic income is the best way to go about it. I fear it’s not a great social arrangement. People should do something to get an income. That’s why I like expanding the Earned Income Tax Credit. We should make it available more broadly to adults without children. And also make vocational training and apprenticeships available after high school.
As technology and automation advance, I have confidence that we have a good opportunity to become more productive. I don’t have confidence that we’ll take advantage of it. Our institutions are being unwound at the moment we need them.
Bianca Vázquez Toness is a journalist based in Cambridge, Massachusetts, who writes about technology and its impact on people. A 2016-17 Knight Science Journalism fellow, she most recently covered India’s technology and telecommunications industries for Bloomberg News.
It’s thelittle changes that make the biggest changes
“But it’s not as if there’s a fixed amount of work to do.”
This is a red herring. Unemployment can increase without there being “a fixed amount of work to do.” Inequality can increase without there being “a fixed amount of work to do.”
People introduce red herrings when they don’t have a clear answer to a question. Stop it. Admit you don’t know.