On The Road, Looking For Robot-Proof Jobs

Apr 14, 2017

A machine from Carnegie Mellon's Biorobotics Lab.
Credit Marketplace/Carnegie Mellon's Biorobotics Lab

Marketplace Morning Report is wrapping up a three-week series, “Robot-Proof Jobs.” Host David Brancaccio hit the road for the stories talking to workers and experts from Pittsburgh and New Castle, Pa. to Milwaukee, Wis. The reporting also drew on the findings of The McKinsey Global Institute, which analyzed the work activities of more than 800 occupations in the U.S. to determine what percentage of a job could be automated using current technology.

Brancaccio spoke with 90.5 WESA’s Larkin Page-Jacobs about what he learned on the road.

Their conversation has been edited for length and clarity. 

PAGE-JACOBS: So David, you’re on this road trip in search of robot proof careers, and I’ve been listening and in one of the stories you say that people who build robots or teach about robotics are the least likely to be replaced by them, but that’s not very reassuring to those of us who don’t program robots. So what are some of the characteristics that make a job robot-proof?

BRANCACCIO: Well that’s what I found the most compelling about the possibility of driving from your fair city to Milwaukee on this road trip. Which is there are lots of jobs if you're blessed with math or science ability—there are the robot jobs, but there's a lot of other jobs that are especially robot proof. The highly skilled jobs are often very immune to the march of technology. But let's say you're artistic, creative, a graphic designer. The robotized portion is fairly low—in the low 20 percent range—because if you're doing graphic design you can't just be cutting and pasting. It's supposed to be innovative, it's supposed to be eye-catching or approach the idea of communicating visually in a new and interesting way. New and interesting is not something that machines are particularly good at. Health care fields, nurses. We talked to a wonderful person named Kay McGee in Chicago on this road trip.  And she talked to me about the deep wells of empathy that are necessary to do her kind of work—something a machine would be terrible at. The ability to read to people. So there are plenty of jobs out there you don't just have to be a robot scientist.

A lot of our work is in controlled environments in a controlled environment. Technology's pretty good at that. A lot of our work that we do in America is data collection or data processing at some level. Technology is especially good at that. This was a calculation McKinsey did that floored me. If you just add up work in controlled environments—data collecting or processing—that's $2.7 trillion dollars in wages. Okay, that's a lot of money, so some business facing publications saw the McKinsey data earlier this year, and said, 'Wow, look at all the money we can save on wages.' Now, that's money that comes out of workers' pockets. And so this is clearly a challenge and an opportunity for society.

PAGE-JACOBS: Besides having certain qualities that a robot can’t have, maybe you have to be agile or you have to have empathy – there are other factors beyond that that would keep some jobs safe from a robot takeover, right? I’m thinking of the cost of building that robot, installing it, that those things may be so expensive that it’s not worth creating the robot that would replace that job.

BRANCACCIO: And in fact the experts point this out. Just because something could be turned into a machine doesn't mean we will. Because if a human is cheaper, then the human always wins. So you have to look at the difference between, well, if we really wanted to spend a lot of money we could get the machine to do this or the algorithm or the software, versus does it really make economic sense? And that saves a lot of jobs.

Here's another one. Social acceptability. There are certain things I think we may not put up with. For instance would you accept a verdict from a robot judge? Robocop? I'm not sure we're ready for that either.

PAGE-JACOBS: OK. Let’s say you had a robot judge. On the one hand you could have a completely unbiased machine deciding a case. They wouldn’t bring any baggage to the table, they wouldn’t have pre-conceived notions like a human might. I imagine that could be a good thing. On the other, they can’t read someone’s facial expressions or tone of voice that might help them make a decision. I can see trade-offs in that. I can see why it might not be politically popular – but I can see an upside. 

BRANCACCIO: Well, certainly and you know you're right on to this important point, which is human judges have all sorts of biases and you could engineer into these things if they were engineered correctly to avoid that kind of bias. It is fascinating. I remember seeing a famous behavioral economics study of parole judges in Israel. Apparently the judges tended to issue very strict rulings would not grant parole when they are “hangry” late in the morning or late in the afternoon when they're getting “hangry.” But after they had had lunch they were much more likely to say yes to a parole. Now there clearly is no legal basis and a machine might be able to do better than that. But still, I mean we're supposed to be you know a jury of your peers, what about the judge?

There's a lot of things that we want to see a human do. It is certainly possible that airliners could be completely robot. It's not technically impossible. They take off and land by themselves, even now. Are you're going to get on that airliner? Maybe someday. So this is a dialog hopefully they'll be having.

But there's so much more and more work will be done by machines. The question is this: Are better, cooler higher paid jobs being created at the same time? So we're going to be watching this.

PAGE-JACOBS: On this road trip, are there any questions that you’ve been left with that you might want to explore going forward?

BRANCACCIO: I think this is going to be a beat going forward. I think that this is not something you do discreetly. We have these three podcast special podcasts that are part of the Marketplace Morning Report regular podcast feed that you can listen to, in which I go into a lot of this stuff in greater depth. We're working on part three now right now. And I've heard from some really smart people some wild policy ideas for dealing with this stuff like "universal basic income." There's this idea of widening the idea of stock ownership in tech companies. We could essentially own the robots and spread the wealth that way if our wages were going down because of technology. There's a lot of other ideas that I just found fascinating and so that's what I want to continue to explore. It's an ongoing story I would say.