At a coastal California redoubt 90 miles south of Silicon Valley, Andrew McAfee told a group of tech luminaries in January that the war between robots and humans, long anticipated by science-fiction novelists, has already begun in the American heartland.
The war is for jobs, he explained, and the robots are winning.
Mr. McAfee, a research scientist at the Massachusetts Institute of Technology’s Sloan School of Management, has been talking about the political ramifications of automated labor for years. He is a co-author of two books on the topic, Race Against the Machine and The Second Machine Age, and has been a regular at conferences where prosperous brainiacs discuss the future.
Over the past year, his admonitions have become more urgent.
"If current trends continue, the people will rise up before the machines do," Mr. McAfee warned at his January presentation. Sure enough, Donald J. Trump won the presidency last year by promising jobs to many of the humans who voted for him.
In the meantime, the MIT researcher is skeptical of ideas like universal basic income, a government stipend that some of his peers suggest could be a stopgap against the poverty and hopelessness that have ravaged formerly middle-class workers.
Mr. McAfee talked with The Chronicle about how the robot-human war might unfold in the coming years, and the role higher education has to play in it.
The political rhetoric about the middle class has been the same for a while, but the composition of the middle class has changed, and a lot of that change has had to do with technology and automation. What middle-class jobs are robots currently poised to take over for good?
The best way to answer that question is to look at the recent trajectory. The recent trajectory has been that technology has been automating lots of routine work. I mean both routine physical work (think about an assembly-line worker in a factory) and routine cognitive work (think about the payroll clerk in that same factory). There are a lot fewer of both of those jobs than there used to be, even though we still actually have a lot of factories in America and we turn out a lot of manufactured goods.
By far the most likely scenario is that that is going to continue and that is going to expand. It used to be the case that if you wanted to listen to another human being talk, figure out what they wanted, and satisfy their request, you had to have a human being involved in that work. That is not true anymore. I can easily envision lots of customer-service type jobs or customer-interaction type jobs that are going to be automated away at the same time. Jobs that don't look routine, but really are, are going to confront automation.
You mentioned customer-service jobs. There are other jobs you’ve mentioned as being at risk of being taken over: clerks, warehouse workers, cab drivers. …
Cab drivers is a little tough. One-hundred-percent-automated cars are going to happen, but they're not going to happen in the immediate future. I think it's going to be a human job for a while yet. But I will say anything more than a decade out and all bets are off.
Here are some other middle-class jobs: associate professor, librarian, curriculum developer, learning technologist. How will the humans in those jobs fare in the long term as robots begin to displace middle-class labor generally?
We are learning that there are viable and effective alternatives to the classic way to acquire higher-level skills that are valuable in the work force, which is to go to a four-year college and get a degree. We're learning, with things like coding boot camps and online courseware, that there are alternatives to that. And I honestly believe that the main thing that keeps those alternatives from spreading even faster is the fact that most employers, for lots and lots of jobs, require the candidate to have a good, old-fashioned degree before they'll take a serious look at them. If and when that changes, I think higher education will change a great deal.
Colleges like to say that they can retrain underemployed adults and also prepare traditional college students to get good jobs in a modern, albeit robot-filled, economy. Can they promise that?
Some of them do, some of them don't. But what I find more interesting is the fact that we're getting pretty compelling evidence that traditional degree programs are not the only way to either provide people skills at the start of their career or in the middle of their career.
People like to talk about the benefits of higher education in terms of job training. They want college to actually prepare people for the work force. I'm curious about whether automation makes teaching to the demands of the workplace like aiming at a moving target.
There is some of that going on. The world is changing, and providing people with the right skills in general becomes a bigger challenge. But honestly when it comes to higher education, the bigger problem that has surfaced over the past decade has been this triple play of the Great Recession, increasing globalization, and higher rates of tech progress. Those three things, I believe, have combined to show us that higher education was not actually working as well as a lot of us thought it was.
You have named educational reform as being among the few things that economists, technologists, and politicians agree is needed. But you have also said that the educational reform that’s been tried has not been sufficient. What kind of reform would be sufficient?
Our educational system is well suited to turn out the kinds of workers the economy needed 50 years ago: those that could read, write, and do some math, and also were trained to follow the voice of authority. Computers are much better than us at math, are learning to read and write very quickly, and are unbeatable at following instructions consistently.
We need an educational system now that excels at producing people to do the things that computers can’t do: figure out what problem to tackle next, work as part of a team to solve it, and have compassion for others and the ability to coordinate, motivate, persuade, and negotiate.
Let's talk a bit about universal basic income. Does UBI mean that, in the future, we all get to practice the liberal arts with impunity rather than compromising and getting jobs in sales and marketing?
No. A UBI big enough and broad enough to cover people that they don't need to go get a market income would be ruinously expensive, even for an economy as big as America’s. A UBI that would provide something like a poverty-line-level income to all American families would take about three quarters of all federal-income-tax receipts. But that’s not going to happen.
Yeah, no kidding.
One of the most important things a society needs to do is not to bestow money on everybody so that they can go pursue whatever it is — a liberal-arts-education life free of the demands of earning a living (for my Silicon Valley friends I say, "all Burning Man, all the time"). But that's not the goal.
When I look at the communities in America where work has vanished the most, these are really troubled communities — everything from single-parent homes, to high crime rates, to mortality increases that are at the same level as the AIDS epidemic. The thought experiment I always conduct for myself is to ask which of those problems would be solved if a check showed up from the government every month for, I don't know, a thousand dollars or something. My answer is, absolutely none of those problems.
I'm a pretty vocal opponent of the UBI for that reason. When you look at the actual data about job growth there is no evidence that we are at peak jobs in the United States. I'll start to take the idea of UBI a lot more seriously when our economy starts growing and overall jobs start shrinking.
I want to address the elephant in the room. Donald Trump says that he's going to bring back great jobs and wages and to restore the middle class. Does he have the ability to do that?
Presidents have more limited ability to influence the economy than they think and than they talk about when they're trying to get elected. Neither Donald Trump nor anybody else is going to bring back an American middle class full of people doing industrial-era routine work. Those jobs are gone to Shenzhen and Bangalore, and they are not coming back. The products that those factories turned out would be massively uncompetitive.
So is there any way that the sophistication of robotics can put the haves and have-nots of the middle class back together again, or is automation just going to continue driving them apart?
There's a great quote from the first person that won the Nobel in economics, who said that inequality is a race between technology and education. Technology tends to increase inequality, and education tends to decrease it. Education needs to do what it's done successfully in previous chapters of the country's history, which is to be a great leveler.
I do think the forces of tech progress are economically polarizing. That doesn't mean that we should just throw up our hands and walk away from the idea of trying to restore a large, stable, prosperous middle class in America. What I think we need to do is give the forces of entrepreneurship and innovation and economic growth and job creation the best possible chance to do what they've been doing very successfully for a long, long time, which is to create lots of demand for labor and lots of jobs, and then good wages may get better over time.
This interview has been edited for length and clarity.
Steve Kolowich writes about writes about ordinary people in extraordinary times, and extraordinary people in ordinary times. Follow him on Twitter @stevekolowich, or write to him at firstname.lastname@example.org.