Robot-Proof Jobs

Robot-Proof Jobs 1: Transcript

Marketplace Staff Apr 10, 2017

Five years ago, Marketplace explored how machines, robots and software algorithms were increasingly entering the workforce in our series “Robots Ate My Job.” Now, we’re looking at what humans can do about it with a new journey to find robot-proof jobs.


David Brancaccio: We are going to be machines soon. I’m David Brancaccio, host of the Marketplace Morning Report. I didn’t write that opening line. It was crafted by a man working the tollbooth on an entrance ramp to Interstate 376 in Pennsylvania.We didn’t have an E-Z Pass for the toll and said it’s nice to see a real person in the booth.

Toll Taker: We are going to be machines soon.

David Brancaccio: Oh no.

Toll Taker: Yeah, they’ll do it, because everybody else is.

David Brancaccio: We are going to be machines soon. At another set of tolls, word from the booth was that person’s position would go in just another month, that would be it. This year in and around New York City, all the MTA bridges and tunnels are going cashless, a euphemism for humanless.

Toll Taker: It’s alright. It was a good job while it lasted. Bye now.

David Brancaccio: Hang in there, sir.

David Brancaccio: Technology is taking its toll on the toll takers. Yeah, well, so what, economists tend to say. No matter that toll-takers like these jobs, need the money and the benefits. Classic economic theory happily acknowledges that technology eradicates loads of jobs like this. Will we ever forget that 98 percent of all farming jobs were annihilated by machines over the last two centuries? We’re taught that’s okay because technology creates more productive, higher-paying, less back-breaking and more interesting jobs. The farm workers got better work in factories and American standards of living went up. Even if the toll-takers don’t get the better jobs as the machines hit, maybe their kids will, fixing the machines, designing them. That’s the classic promise, anyway.

Then, the world changed. This March, a pair of respected researchers, one from MIT the other from Boston University, published a working paper for the National Bureau of Economic Research. Using real data from the real economy, not conceptual models or anecdotes, their results strongly suggested the more robots there are in factories, the fewer human jobs there are in general. Human wages went down, too. This new paper has been greeted in economics with a stunned, “huh.” It’s like when Copernicus showed his early evidence that it’s the sun in the center, not the planets.  

Robots are just one kind of technology that competes in the workplace. We can talk about machines, software, algorithms, artificial intelligence, as well as robots on a factory line. And this new academic paper, bombshell that it is, doesn’t settle the argument about what tech does to jobs. But it’s one hell of a news peg, as we say in my line of work, on a topic with deeply human dimensions.

We’re calling this Robot-Proof Jobs. In this set of three special shows, I want to look into what we can do about all this. Which jobs stand the best chances of survival as more technology comes into the workforce?  How do we train people for these jobs? And I’m going to use my one, personal superpower the ability to get really smart people to answer my phone calls to explore some wild ideas for re-engineering the economy to give people a chance to maintain a decent livelihood.  People like this Marketplace listener in Utah.

Christina Summers: So I’m Christina Summers, and I started out as a real estate appraiser, and I actually got pretty good and I specialized in the Denver in the higher-ed homes, and I made pretty good money doing it.

David Brancaccio: A house is the big collateral for a mortgage, right? Banks want to know what it’s really worth. Summers would visit the house, poke around, look up information on the property, check for zoning and other compliance issues. Until seven or eight years ago, a pile of new technology falls from the sky. This included Google Earth — just zoom in and look from a satellite. Plus a new algorithm, something called an AVM.

Christina Summers: An automated evaluation model. So, basically all of this information is public data, and so what you do is you create a computer program that mines all that public data and they can pull the data just looking at valuation and looking at what other homes in the area have sold for. So, the appraiser now is just superfluous, it’s just not necessary anymore.

David Brancaccio: Summers tried to go back to school but found the math and science they never taught her in  high school would add years to the degree. In her late thirties, Christina Summers found work.

Christina Summers: Right now, I am just driving a bus.

David Brancaccio: For the city of Ogden, UT. Doing home appraisals, Christina got paid $65,000. After the technology descended, the offers she got had fallen to between $22,000 and $26,000 a year.

Christina Summers: Driving a bus, I get paid $40,000 a year for that. So the schedule is brutal, there are a lot of brutal things about it, and it’s quite frankly humiliating to go from all these dreams — I mean, they push you so hard in college — and now, I’m a bus driver. And to have to face that is always difficult. But there are good things about being a bus driver, and I will always persist on finding the good things.

David Brancaccio: She’s just hoping the city doesn’t someday opt for self-driving buses, which might be cheaper.  

Christina Summers: Not only that, but computers don’t have to take bathroom breaks and they can run all night. I only need ten years, that’s all I ask. Just give me ten. I’ll be OK with that.

David Brancaccio: Look, bus driving is honorable work. But you can hear how technology is causing wrenching changes for people — the vanishing toll booth attendant, the vanishing army that used to come round to read gas, electric and water meters. I’m here to say there are probably about 160 million reasons to care about Christina Summers’ story. 160 million is the size of the American labor force, not even counting military.

Because here’s the thing:  It’s not just the obvious jobs that the laser beams of technology are melting into nothing. The machines and the software may be coming after your job and mine.  And what about your offspring, if you have them or ever will have them, what kind of careers will they have in the face of the robot competition?  Again, many experts say it’ll all work out in the end. We’ll get richer. Maybe machines will do just about everything and we won’t have to work, and the jobs we do will be way more interesting.  Maybe.

Martin Ford: When it comes to technology and machines, we’ve always got kind of an optimism bias. We always assume absolutely that technology is going to be a good thing for us, and historically it has.

David Brancaccio: They key word there is “assume.” We’ll get back to that guy later, that’s Martin Ford who wrote a book with the subtitle “Technology and the Threat of a Jobless Future.”

Another reason I wanted to look into this whole topic now is that technological unemployment, as it’s called, may be one of the biggest reasons people are so hacked off in America these days. I think it’s behind a lot of the anger that defines our relationships and our politics at this moment in history.  When workers don’t reap the benefits of automation, it can amplify income inequality, hollowing out the middle class and make people desperate. This, even as our attention is focused elsewhere.

President Trump montage: …our jobs are being taken away from us… // …shipped thousands and thousands of miles away… // …to Mexico… // …going to other lands… // …and overseas… // …and we’re going to keep our jobs in our country…

David Brancaccio: Well, guess what. A Nobel Prize-winning economist has told me flatly that most of the American jobs going overseas have already gone. And the call for bringing overseas jobs back to our shores? No complaint on that from me, but here’s the fear: the factories that might come back from abroad won’t hire like they used to, not even close because of machines. I once did a story about a wonderful guy who pulled Ingersoll Machine Tools out of bankruptcy in Rockford, Illinois. The firm’s very successful now and has 250 employees. But back in the day, it employed not 250, but 2,000 people.  

And, if you want to really torture yourself, search the phrase “lights out factory.” Lights out manufacturing with machines only, no humans included. In the dark because what do machines need with overhead lighting?  Kurt Vonnegut’s dystopian novel Player Piano is all about this. Did you ever see the full title of the first edition?  “America in the Coming Age of Electronics,” he called it back in 1952.

I’ve been watching this evolve over time. Five years ago, I tried to illustrate for radio listeners how human-free our tech-fueled economy had become. I was able to drive from the Atlantic Ocean to the Pacific Ocean: six days, five nights. In teched-up America, it was like a solo mission to Mars. I never once transacted business with a human being — checking into hotels with a robot check in clerk, using a kiosk that spits out my room key, E-Z Pass at the tolls, a swipe at the gas pump, scanned barcodes at the self-checkout for my food. Over land, sea to shining sea, without dealing with a person.  At grocery in Roanoke, VA, the maven of the self-checkout lanes really wanted to help me scan my ear of corn, so I had to run for it.

 David Brancaccio: No receipt?

 Clerk: I can take that.

 David Brancaccio: Whoops!

David Brancaccio: Now, that trip we called Robots Ate My Job. It’s five years later now, I wanted to get more practical and forward-looking. This time, I’m looking for jobs that either can’t easily be replaced by technology, or jobs get along well with tech. I’ve chosen a section of the country for this new road trip that is loudly on record in the last election as clamoring for change. I’m driving across an arc of the Midwest from Pennsylvania through to Wisconsin. The title this time is Robot-Proof Jobs, or how I learned to stop worrying and to love the bots.


Road Trip Chapter 1: Pennsylvania

David Brancaccio: Let’s get this road trip underway. At the car rental counter in Pittsburgh, I request something domestic, so they hand me the keys to a Nissan Altima, most likely assembled in Smyrna, Tennessee or Canton, Mississippi, a plant with 6,400 humans and 1,200 robots. Sharing the driving duties, Marketplace producer Katie Long.

Katie Long: Hi.

David Brancaccio: A person of few words. And a time-tested road trip companion: the car radio.

Car radio: …Gino Stone to the basket, 40 to 32, New Castle…

David Brancaccio: The basketball broadcast in New Castle, Pennsylvania, an hour outside of Pittsburgh, where we find Linda Spinelli’s house.

Linda Spinelli: How are you? Come on in.

David Brancaccio: Linda has one of those jobs that robots, software, and algorithms will have a tough time taking away.  Linda does HR. She’s a human resources manager for a company here with about 200 employees.

Linda Spinelli: And actually we are growing. I’m in the process of trying to hire 15.

David Brancaccio: Hiring, firing, health plans, retirement stuff. It’s an HR department of:

Linda Spinelli: I’m the human resources department. One. Me.

David Brancaccio: So, it’s up to you, Linda.

Linda Spinelli: One. But, this company, you get to know the people, you get to know a lot about them, personal stuff. For example, a person falls asleep at the job. You would think that would be immediate termination, but it’s not. What you do, is you need to go up to them and then gently wake them up and then kind of question them. Why did you fall asleep? Were you playing videogames last night? Or, maybe there is a health problem. Not a robot going up and waking them up.

David Brancaccio: The consulting firm McKinsey has a research division that’s looked at just about every job they could think of in America. They’ve also counted how many positions there are in a given field in America, what they pay, and here’s one innovation what slice, what percentage of any job is either done by machines now or could be done by machines or software, using current technology.
HR professional, like Linda Spinelli here in Pennsylvania, is just one of a range of middle-management jobs that are relatively low in repetitive, rote work and relatively high in hands on interpersonal work, making them less vulnerable to replacement by technology. This is why we’ve come out here to meet Linda, to understand the qualities of a job that won’t get easily replaced by a machine. I’ve been holding back on the kind work Linda’s company does. Here’s a hint:

[Fireworks sounds]

David Brancaccio: New Castle, PA calls itself the fireworks capital of America. Two of the biggest pyrotechnic companies got their start here. Linda’s company doesn’t do 4th of July, but it does use explosives, to make holes in the earth.

Linda Spinelli: They do above and below ground coal mining. We help with utilities when they’re laying pipe or whatever, there might be rock in the way. 

David Brancaccio: How does that McKinsey automation study rate those Linda Spinelli-style human resource manager jobs? Producer Katie’s got the scorecard with her.

Katie Long: Indeed, I do. The portion that’s automatable, or robotizable, is 22 percent, which is on the low-end. Average pay is about $58,000 a year. And, there are plenty of HR jobs nationwide, more than 400,000.

David Brancaccio: So, machines could do less than a quarter of the work of an HR manager, around the same robot potential, it turns out as, for instance, a lawyer. That’s a pretty small slice that could be done by machine.

Linda Spinelli: I really don’t see it as a threat.

David Brancaccio: Contrast that to, for instance, the job of pharmacist where machines and software could do nearly half the work. One of my colleagues just came back from Australia where a human checked his doctor’s prescription, but all the rest, the finding of the pills, the packaging, labeling and dispensing, was done by a pharmacy robot.

A frequently cited Oxford University study in 2013 looked at computerization of jobs and found that nearly half, 47 percent, of jobs in America could be automated in the next 20 years. What the McKinsey folks have now done is come out with a more finely-diced way to look at all this. Since we’re relying on some of that data for this trip, I made sure to connect with one of the lead authors, James Manyika, director of that McKinsey Institute.

James Manyika: We started with the idea that anybody’s job or any occupation consists of 10 to 20 different activities or tasks.

David Brancaccio: Some of Linda Spinelli’s HR work uses technology, the internet. But a good deal she carries out in person — judging character during the hiring process, or physically escorting potential new drivers of trucks loaded with explosives to and from the restroom for the legally-required drug test. The McKinsey researchers first classified the different tasks that make up a job. Then, part two: researchers asked which slices can software, machines, algorithms, robots perform instead, at least in theory.  

James Manyika: So, the collection of data, the processing of data, working in very predictable physical environments. Those three kinds of activities alone make up something like nearly half the activities in, say, the U.S. economy if you measure activity levels. And, you know, we pay in wages something like $2.7 trillion in wages to those three kinds of activities. And those are the most highly susceptible to automation.

David Brancaccio: When McKinsey published this study in early 2017, some management-facing publications put all those trillions of dollars in wages into breathless headlines about all the money industry can save, $2.7 trillion, exclamation point. Shareholders, rejoice. For workers, that’s money right out of their pocket. But the McKinsey team went beyond deciding if machines could do the work using present technology. They also considered the cost. If a human’s cheaper, than the human wins.

James Manyika:  So in some ways, if the question was simply about which jobs are safe, I would actually argue a lot of low-skilled physical work in a highly unstructured environments is actually relatively safe. Versus, for example, a lot of what historically has been highly-skilled cognitive, but highly automatable tasks.


Road Trip Chapter 2: Ohio

David Brancaccio: Back on the road trip, where we’ve moved into Ohio. This college radio station is on a roll.

[Car radio music]

David Brancaccio: A factor to consider when evaluating how jobs will be robotized is social acceptability. An artificial intelligence is getting tested in Europe that might some day act as judge. Hundreds of previous cases from the European Court of Human Rights were fed in and the software picked the right outcome in 79 percent of cases. Not too shabby. But experts question whether society would accept a robot judge. Egads, you might say, but what if it’s not a criminal case, maybe an arbitration?  A machine that’s decent at picking legal outcomes is one example of this crazy era in which we now live when it comes to technology and the labor force.

Car radio: Hello folks, this is WOBC, Oberlin 91.5, Oberlin College radio…

David Brancaccio: Oberlin, Ohio. I came to the famous conservatory here to meet a guy with a job which, the data suggest, is immune to the robots. Tom Lopez is a composer.


David Brancaccio: A choreographer in California had sent a video of a performance in rehearsal. Tom’s job: create the score from the cues in this video.  

Tom Lopez: That yellow, sort of  rectangular blob is a couch. And it’s strange that it’s tipped over like that, so why did that happen?

David Brancaccio: Tom, a human composer, not some algorithm, parses some of the stage props in the video. There’s that couch. A dancer wears a yellow kitchen glove, which the composer connects to a theme of household drudgery, which inspires his score.


David Brancaccio: The fanciest computer visualization system on the planet will not, I repeat not, make connections like that.  

Tom Lopez: I like working with patterns, but what makes music interesting and attractive to most audiences are the slight variations in the pattern. It’s where you break the pattern and break the mold.

David Brancaccio: And decide whether that new sound is any good. Computers are completely clueless at judging artistic merit. That hasn’t stopped some scientists from creating an artificial intelligence that hasn’t composed some music.

[Artificial intelligence generated music]

David Brancaccio: Eh. And even if computers end up composing muzak-style schmaltz for customers at Walgreen’s or CVS, technology isn’t taking away composer jobs. Tom says it’s creating more of them.

Tom Lopez: There are new opportunities in composing and music technology that didn’t exist ten years ago. Everything online, computer games.

David Brancaccio: Computer games need music, of course.

Tom Lopez: So, we’ll let muzak go.

David Brancaccio: This is one of the strongest arguments for those who hope the fire hose blast of technology hitting the workforce now will all work out for the best. Tech can create further opportunity, for those with the skills and education to take advantage.

Tom Lopez, in keeping with our theme here of occupations that play nicely with technology, does some styling on a vintage synthesizer, an ARP 2600.

[ARP 2600 synthesizer]

David Brancaccio: That is the synthesizer model used to make the sounds of R2D2 in the movie.  Katie, the McKinsey numbers?

Katie Long: Let’s see here. Music directors and composers, zero percent automation potential, pay $52,000 a year. But there aren’t that many of them, only 22,000.

[Car sounds]

David Brancaccio: Less than an hour from Oberlin, Sandusky, Ohio.

[Police siren]

David Brancaccio: No, we didn’t get pulled over. Here’s a job where they’re hiring nationally.

Bronson Lillo: I’m officer Bronson Lillo. I’m a police officer with the city of Sandusky in Ohio.

David Brancaccio: A ride-along, where I chose to get imprisoned in the back with the rock hard seat, barred windows and door with no escape.  

David Brancaccio: You can leave me in here forever, it’d be awesome.

Bronson Lillo: When I go up to the car, you know, yes, we have our way that we’re supposed to talk to people. The things we’re supposed to say to you, to get that traffic stop moving and get it over with in a timely fashion, and get the people back on their way. But sometimes when you talk to people, maybe you know, they just experienced a loss in their family or maybe something serious is going on and they’re, you know, going a little above the speed limit because they need to get back to their house and make sure everyone’s OK. Like there’s a time and a place to write a ticket and there’s a time and a place to, you know basically, take a backseat and let that person talk through their problem or what’s going on.

Katie Long: Police and Sheriff’s patrol officers, machines could only do about 19 percent of the tasks they do, which again is pretty low compared to all other jobs in the US. The pay is pretty good, and there’s a lot of them,  635,000.

[Police radio]

David Brancaccio: Again, plenty of tech in this robot-proof job here. Officer Lillo’s cigarette pack-sized body camera, his Mobile Data Terminal, a ruggedized laptop in the car. But a robot cop isn’t in our near future. Even if the tech ever made a robo-cop possible, too much automation in law enforcement may not pass the social acceptability test. On anti-riot duty some day? Maybe. Soldier versions, way before civilian uses. But, before we get too science-fictiony here, how about a reality check?  


Road Trip Chapter 3: Robot Auditions

David Brancaccio: In preparation for this trip, I had wondered about bringing along a robot sidekick.  After all,  Kerouac’s character had Dean Moriarty as a sidekick. On that solo trip I did for Marketplace five years ago, I had figured out how to hack my GPS-navigation unit to put in my wife Mary’s voice.

Mary Brancaccio: Turn the second right, gorgeous. Turn the third left, gorgeous.  

David Brancaccio: So I didn’t feel so lonely on all those miles by myself behind the wheel. For the journey this time, I had wondered if I could find a robot Sancho Panza to my Don Quixote. And for that, I held a series of auditions for some state-of-the-art road trip companion of the hardware or software kind. It’s one test of how far tech has come in automating more than just pieces of humans, but whole humans. Plus, a robot journey, like all good journalism, needs some balance. You know, get the robot point of view. We’ll meet a few of the contenders throughout this series, starting in a high rise in downtown San Francisco.

Pepper: Can I help you find a store within our shopping center? Oh, are you looking for Victoria’s Secret?

David Brancaccio: Yes.

Pepper: Perfect, here it is.

David Brancaccio: Pepper, the robot, looks to me like a life-size statue of a precocious 9-year-old. It’s big, Manga-style eyes are ringed with flashing LED’s, and there’s a touchscreen molded to its chest.  The head speaks but the mouth never moves. It’s Pepper’s posture that’s most expressive. It’s hinged at the arms and fingers, and bendy at the neck and waist.

Steve Carlin: Pepper is there to engage with you for as long as you need and can answer any question multiple times, will never get upset, will never get frustrated, will never get impatient.

David Brancaccio: Decent qualities for an extended road trip. That’s Steve Carlin of Softbank Robotics America by the way. If Pepper is designed as any kind of a buddy, it’s a shopping buddy. It can answer questions when you’re lost in the mall. It has helped in car showrooms. There’s one working at the Oakland Airport, pushing drinks.

Pepper: Let’s go for a medium-bodied. The Pyramid India Pale Ale is a medium-bodied light golden ale.

David Brancaccio: My kind of robot. Pepper learns from experience, using artificial intelligence derived from IBM’s Watson. In Japan, where the first thousand units of these sold out in less than a minute, there’s precedent for Pepper as companion.

Steve Carlin: There was a short documentary done on this woman who literally will take Pepper around town with her, she’s built a cart. She actually had a really strong relationship with her grandfather who recently passed away, so she took Pepper to the cemetery to introduce Pepper to her grandfather.

Pepper: Oh, by the way, what do you think of my presentation? Give me a rating from one to four.

David Brancaccio: Thing is, while it bends in 17 different ways and scoots around on rollers, Pepper can’t sit, which makes it tough to wedge into a car seat for a road trip. So onto the next robot audition, this time in  Redwood City, California. He-she-it is called Kuri, K-U-R-I.

Sarah Osentoski: Kuri is an adorable robot who looks a little bit like an intergalactic space penguin.

David Brancaccio: Kuri is knee high with a nodding and swiveling head, big dark eyes and no nose and mouth. The arms are just suggestions, permanently glued to its sides.

Sarah Osentoski: And she blinks and she turns her head, and she pays attention to things.  

David Brancaccio: That’s Sarah Osentoski, engineer and chief operating officer at Mayfield Robotics, owned by the German company Bosch. Kuri is equipped with facial recognition and can travel around the house beaming video to you remotely if you’re away.  It can play back music, audio books.

Sarah Osentoski: She can be an audio companion, so she can play NPR for you in the morning when you’re getting ready and follow you throughout the house.

David Brancaccio: Never escape the Marketplace Morning Report with Kuri dogging your heels. Also, there’s an app that turns Kuri into a cheeky mobile alarm clock that sneaks into the sleepy kids room, sidles up to the bed and goes:

Kuri: Ahhhhhhhhh!

David Brancaccio: Keep you awake on a road trip, I suppose. And it would fit in a booster seat next to me. But here’s the key thing about Kuri. It’s not going to replace a human and is actively engineered to remind me of its lack of sophistication.

Osentoski: When you start to have something that drives around your house and expresses emotions, you have a lot of expectations about how smart she might be. So we want to set expectations that Kuri is more like, maybe a medium smart dog at best, and less like a person, like a kid.

David Brancaccio: Alas, Kuri won’t be ready until the holidays this year, so no road trip for this robot.  I’ll hold more auditions for automated sidekick later.


Road Trip Chapter 4: Michigan

David Brancaccio: Back to the road trip to find more robot-proof jobs. Southeast Michigan now.

[School bell music]

David Brancaccio: Wow, you really can’t teach over this, you’re class is over when this happens.

David Brancaccio: That’s not the radio, that’s how a 21st century school eases students from one class to the next. No jarring bell or buzzer at Madison High School in the town of Adrian, in Michigan. I’m back looking for jobs with human qualities technology won’t soon replace. Angela Tedora is confident she’s a pretty robot-proof English teacher.

Angela Tedora: That fear is not something that I worry about because there are certain things that we do in the classroom that cannot be replaced by a machine.

David Brancaccio: Robots, schmobots.

Angela Tedora: There’s so much more that goes into it than just being able to have a student load a learning management system, and answering questions, or going through activities in that regard.

David Brancaccio: What’s the data say, Katie?

Katie Long: Machines can do about 20 percent of what secondary school teachers do. Pay, $54,000 a year, but there are a lot of them, 945,000 in America.

David Brancaccio: While some school districts hand down curriculum, work at the front of the class is very seldom from a script or recited rote in the way a machine would like it.  A teacher needs wells of empathy and the knack for detecting when the material is or is not connecting. Ms. Tedora has her seniors reading a novel called Ready Player One in which the earth’s a mess in 2044 and everyone attends virtual school from home using 3D goggles. This does remind us of those MOOC’s, the Massive Open Online Courses, universities are pushing. For high school, maybe they’d work for some subjects, for the most motivated of students.

David Brancaccio: But the thing is, day to day, as a teacher you need to be physically present to see if everything’s OK, or to set somebody straight who falls out of line, right?

Angela Tedora: Absolutely. Absolutely, by having those role models in their lives, that’s extremely important as well.

David Brancaccio: Deeper into the future, who knows? One expert talked to me about a more distant future in which learning is immersive through technology and what used to be teachers are more like exam proctors or hall monitors.  

Angela Tedora: We just had some professional development here at school that talked about the jobs that kids at our elementary school level right now, the jobs that we’re preparing them for haven’t even been created yet. And really, that’s a hurdle that we need to figure out how to get over because how do you prepare them for the unknown. And it’s problem solving skills, it’s them being able to solve whatever’s put in front of them.

David Brancaccio: “The jobs we’re preparing them for haven’t even been created yet.” Teach what, exactly? I want to come back to that question before we’re through.

A longer stretch of this road trip west on Interstate 94 from Michigan, clipping Indiana then into Illinois. It’s too early for buds on the trees to open and what you see is monochrome branches set against the late winter sky. This calls to mind a weird detail in the robot-proof jobs data. Some of the low-end jobs you’d think would be vulnerable to technology are actually pretty safe. In the McKinsey study, tree pruner, groomer of trees, only has a robot potential of five percent, if you don’t mind the pay. This is an example of something called Moravec’s Paradox.

Erik Brynjolfsson: This idea that the simple skills can be harder to automate and the very advanced cognitive skills are sometimes easier to put into a machine codeable form.

David Brancaccio: Erik Brynjolfsson is an economics professor at MIT’s Sloan School of Business. He is the author of the best seller The Second Machine Age: Work, Progress and Prosperity in a Time of Brilliant Technologies. The old way, he says, was a fairly straight line

Erik Brynjolfsson: Low-skilled jobs falling in demand, middle-skilled jobs sort of in the middle, and high-skilled jobs increasing in demand. And economists call that “skill bias technical change,” you know, change that’s biased towards the high-skilled groups. More recently, it’s become sort of more of a bowl shape, where people in the middle are really the ones being hammered most. Middle-skilled jobs like technicians, middle managers, travel agents — jobs that require some information processing skill, but they are not, you know, PhD’s or super-star rock stars.

David Brancaccio: We hearing all this? Not just the low-skilled ones are toast. It’s the jobs that are skilled, not with the highest skills, but the middle-skilled ones that are getting — what did Erik say again? They’re the ones…

Erik Brynjolfsson: Being hammered most.

David Brancaccio: Which does this to people earning the typical wage in America.

Erik Brynjolfsson: One of the results has been that median income, the income for the person at the 50th percentile, is lower today that in was in the late 1990’s, about 20 years ago. It’s a remarkable thing that’s never happened before in American history to have wages stagnate for the typical person for that long a period.

David Brancaccio: And here’s the thing, sometimes people went to college, that doesn’t guarantee you a whole lot these days?

Erik Brynjolfsson: It doesn’t guarantee you, but I just want to be clear, certainly people with college educations still significantly outperform people with only a high school or less education. They have less unemployment, but they are not seeing their wages grow the way they used to. And especially sort of the people at the middle or low end of the college educated group, they have been hit hardest. Think of a tax preparer. That was something that requires, you know, a certain bit of skill with numbers  and logic and working through complicated tax law and forms. But increasingly machines are doing that. A lot of us use TurboTax or other sorts of programs to take care of the routine parts, routine information processing, that humans used to be uniquely good at.

David Brancaccio: Did you see who H&R Block just hired to do some of this work?

Erik Brynjolfsson: Yeah, they hired my friend IBM Watson, right?

David Brancaccio: Not just tax preparer. Even some of those fancy jobs on Wall Street are getting eaten alive by this right now. At the investment company Blackrock this spring, it’s out with many skilled investment advisers and in with the algorithms, which run some portfolios very well for lower cost. Remember when the stock market was people, screaming and yelling?

[Stock market open outcry]

David Brancaccio: And what about that massive industry, health care? Radiologists may not want to hear that artificial intelligence is getting pretty adept at reading X-rays, and pathologists have been getting help from machines which do the first pass at many Pap smears. But there are parts of health care that technology will have a hard time conquering.


Road Trip Chapter 5: Illinois

Car radio: This is WBEZ Chicago, 91.5.

David Brancaccio: Back on the road and I have another robot-proof job for us. Two letters, O-T, stands for Occupational Therapist.

Kay McGee: My name is Kay McGee. The whole goal of Occupational Therapy is to help people be as participatory as possible, as independent as possible in their everyday activities, whatever is most important to them.

David Brancaccio: McGee is an occupational therapist at the University of Illinois Hospital & Health Sciences System. “Occupation” is defined as anything a person thinks is important to his or her day-to-day life after an illness or an injury.

Kay McGee: So we’re really looking at enhancing quality of life, but also safety and independence.  

David Brancaccio: It could be eating or typing or relearning public transit for example. One 20-something patient with a brain injury wanted Kay to help her get back two skills right away. Tying a ponytail or shoes, one-handed. Kay showed me.

Kay McGee: Here’s where you use that movement between the two loops.

David Brancaccio: Like this?

Kay McGee: Yep, yep.

David Brancaccio: She uses tech, the shoe-tying strategy came from YouTube. But the center of McGee’s job is listening, teaching, and teamwork. The numbers, Katie?

Katie Long: Occupational Therapists, 25 percent of what they do could be done by machines, which again, is pretty low. The pay is pretty good, $77,000 a year, and there are about 100,000 OT’s in U.S.

Kay McGee: Before we can seek to partner and offer services, we need to really understand what someone is looking for from us.

David Brancaccio: You don’t want to just set them up in front of a touchscreen, have them press a couple buttons and find out that way?

Kay McGee: If you’ve just had a stroke and you’re trying to figure out how to get to the bathroom by yourself or you’re a new parent whose child isn’t able to feed, a touchscreen may not be the best choice. It may not be responsive enough or prepared for your emotional reaction, if you’re talking about end of life goals or new vision loss or acute psychiatric changes that have brought you back into the hospital again.

David Brancaccio: The Bureau of Labor Statistics projects the country will need 27,000 more robot-proof OT’s by the time a high school freshman now gets out of college. That’s a lot in the next four years. Important work, resistant to incursions by technology, crucial contribution to patients in need, what more do you need for a career choice?   


Road Trip Chapter 6: Wisconsin

David Brancaccio: One of the running gags as I’ve delved into all this has been my discovery in the McKinsey numbers is that a very robot-proof and very highly paid job  is, wait for it, CEO.  

Katie Long: The scorecard shows just like OT, a CEO’s job is only 25 percent tech, meaning its robot potential is quite low. What’s high is the pay, $186,000 a year on average.

David Brancaccio: If you thought it was awesome to be a CEO now, it’ll probably stay super fabulous in the future. And not just because the robots will be boosting profits thereby contributing to CEO compensation packages. Part of it is that people follow leaders, not machines. Building consensus may be an especially human talent. So there you have it, be a CEO, for heaven’s sake. MIT’s Erik Brynjolfsson again.

Erik Brynjolfsson: I think that’s right and I laugh a little bit because everyone is like gee, well, we can’t have a bunch of CEO’s. But you know, let’s just push a little harder on that. Actually, the economy is changing in a way where we will have more and more CEO’s. A lot of us are going to be CEO’s of our own one-person business. So you may run your own yoga practice or give take care of kids or take care of older people, do selling. Or for that matter, do creative work, create YouTube videos or video games. We have this long tail of tasks where people are running their own businesses and they are inventing jobs for themselves. They are figuring out what the market demands, they are figuring out how to work with suppliers, basically doing all the tasks of a CEO, and machines can’t do that.

Car radio: Winnebago Land Barbershop Chorus presents “Yeehaw” at the Oshkosh Auditorium…

David Brancaccio: Yep. Gotta love Wisconsin, the final state of our six state road trip.

Michelle D’Attilio: Hi, my name is Michelle D’Attilio. I am CEO of Soch. We are a social media slash corporate branding slash ever evolving agency.

David Brancaccio: Soch, in downtown Milwaukee, helps other firms manage their social media. It used to be, you helped a small company build a fancy website to look big.

Michelle D’Attilio: Today, what we do is we help large businesses seem small, right? So, again, going back to working with their consumer on a one-to-one basis.

David Brancaccio: Michelle’s company, with its platoon of social media experts entangled in a cluster of keyboards and screens, works on behalf of regional and national brands to figure out what happens when customers want to interact using social media. Say somebody tweets at the Briggs & Stratton lawnmower engine company here, asking if gas ever goes bad in a lawnmower. Soch people figure out who answers, how, and ways that Briggs & Stratton can build on that relationship. Soch is one of those “who would have thought of that a dozen years ago” kinds of companies.

Michelle D’Attilio:  It’s I think why entrepreneurs are so very valuable to the country. Because, we, there’s not plan. There’s no book that tells me how to run a social media company, right? I don’t even know, social media didn’t really exist 15-years ago, right? And I think that’s the beauty of entrepreneurs is that we, we’re so used to, OK, this happened, you now have to ricochet off, you have to figure out what the next step is.

David Brancaccio: It’s something the man from McKinsey noticed. James Manyika says with all this technological change, the job category in America that’s growing the fastest is the one marked “other.”

James Manyika: We are always going to be creating new job categories that don’t exist. So to the extent that we can encourage companies to be innovative, to encourage the creation of these new activities, I think that’s always going to create new job categories that we can look forward to.

David Brancaccio: And you’d think it’s entrepreneurs like Michelle D’Attilio causing this “other” job category to blossom, given this time of rapid technological change. You’d think. MIT’s Brynjolfsson again.

Erik Brynjolfsson: I certainly think entrepreneurs will be one of the jobs of the future, but it’s not something that we have enough of. One of the great ironies I discovered in my research was that over the past 20-years we have fewer entrepreneurs and fewer startups that we did in the past. We need a lot more of it, not just because it’s a great job for the people who are doing it, but also because it helps create jobs for other people. But to have more entrepreneurs, we need to reinvent our education. We need a better safety net. We need portable health insurance. We need reduced regulation, less occupational licensing, and we need a change in culture.

David Brancaccio: A change in culture may need new policies. Big new policies will take time.  And some of those policies I am finding out about are doozies, verging on the “what the heck did you just say?” mind-blowing. I promise you, we’ll get to those. But first, in Part Two of this botcast: Why the most robot-proof jobs in the world are a little like playing god.

Note: Transcript may contain typos or errors. Please cross-reference with the corresponding audio to ensure accuracy.

There’s a lot happening in the world.  Through it all, Marketplace is here for you. 

You rely on Marketplace to break down the world’s events and tell you how it affects you in a fact-based, approachable way. We rely on your financial support to keep making that possible. 

Your donation today powers the independent journalism that you rely on. For just $5/month, you can help sustain Marketplace so we can keep reporting on the things that matter to you.