Robots will not take over most jobs

Byron Reese, publisher of Gigaom, author, futurist, and host of the Voices of AI podcast, says robots will not take over most jobs, and discusses narrow v. broad artificial intelligence.

Robots will not take over most jobs

Tonya Hall interviews executives, authors, and thought leaders for our sister site ZDNet, and we're running a selection of some of her most viewed videos. The following is an edited transcript of her conversation with Byron Reese, publisher of Gigaom, author, futurist, and host of the podcast Voices in AI. To watch more of her videos, check out The Tonya Hall Show on ZDNet's YouTube channel.

Tonya Hall: Robots will not take your job. At least that's what my guest says, and he's got some pretty compelling evidence to back it up.

Most people recognize Gigaom as a leading technology research publication, but you've written a book, The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity. Your book begins, actually, with recounting the changes that humans have faced since the beginning of civilization. In fact, you seem to embrace change and are optimistic about it. So how does that relate to the changes pushed upon society today by advances in technology?

Byron Reese: Well, I start off with the idea that the technology is really kind of a profound thing, in that it multiplies human ability. It isn't just kind of a convenience, it really does change us. In the past, there have been, I think, three times new technologies came along which were so profound in their implications, they altered the trajectory of humanity forever. One was 100,000 years ago when we got speech and fire; one was when we got agriculture and the city; another was when we got writing and the wheel.

We're coming up on a fourth one--a time when technology is about to change the course of history. And just like we use writing to kind of outsource our memory, we're using computers to outsource thought and robots to outsource action. And so the question that I wanted to look at is: If you have machines thinking and acting, what then are humans for?

SEE: How iRobot used data science, cloud, and DevOps to design its next-gen smart home robots (cover story PDF) (TechRepublic)

And then I noticed something kind of interesting, which was: Very informed people on these topics have radically different conclusions about what was going to happen. And that really intrigued me, because you would assume that a given room full of experts would kind of have some consensus, but it isn't the case at all. People have these radically different viewpoints about what artificial intelligence is going to do to automation, to humanity, the whole rest.

And what I kind of figured out in writing the book is that it isn't that those people know different things, it's that they believe different things. And I really wanted to understand those core beliefs that drive people's interpretation of the future.

Tonya Hall: There is an Oxford University study that said 45% of jobs will disappear in the next 25 years, in fact, because of robots and artificial intelligence. So that makes many people nervous, but not you. Why not?

Byron Reese: Well, the piece of research you're referencing is high quality and very good, but it doesn't actually say that. That's how it's reported, but the writers added about 600 words that talked about the limitation of the study, and they began by saying, "We're not saying how many jobs are going to be lost. We have no idea." What they actually said, is that 47% of the things you do in your job can be automated, and that actually isn't particularly interesting news. I mean, if you look back 25 years, maybe something like that has been automated in terms of what you do in your daily life.

SEE: Machine automation policy guidelines (Tech Pro Research)

Furthermore, I think if you look across the span of human history, I would say the last 250 years, you have full employment, interestingly. Employment in the US stays between 5% and 10% for 250 years, other than the Depression, yet, you had all this technological change. You had electricity. You had the assembly line. You had the replacement of animal power with steam and all of that. And unemployment never went up.

And I tried to figure out what the half-life of a job is. What's the normal churn of the economy for just destroying old jobs and replacing them with new ones? And I don't have the sense that this is particularly any more disruptive to automation than, say, electricity was to business, or the assembly line. If you were a craftsperson when the assembly line came out, that would seem like a very frightening kind of artificial intelligence. Here's this thing: It makes better goods cheaper, and you don't have to be trained... and what a frightening piece of technology. And yet, unemployment never went up. The standard of living constantly rises and so forth.

So what it turns out happens over the course of time is that technology is really good about creating new good high-paying jobs, and it, in turn, destroys low-paying, lower-skilled jobs. And what people say, the reason they say they're concerned about that is, "Well, do you think those people that get laid off from taking orders at a fast food restaurant are going to become geneticists or these new jobs?" And that isn't what happens at all.

What happens is that a college professor gets the new geneticist job, then a high school biology teacher gets the college job, and a substitute teacher gets hired on, all the way down. So what happens... the question to ask is not whether the people whose jobs are replaced by automation can do the new jobs. The question is, can everybody do a job a little bit harder than the one they're doing now? And if the answer to that is yes, which I firmly believe, then every time technology creates a new job, a new awesome high-paying job, everybody gets a promotion.

SEE: IT jobs in 2020: A leader's guide (ZDNet special report) | Download the report as a PDF (TechRepublic)

Tonya Hall: Is there such a thing then as "robot-proof" jobs?

Byron Reese: I'm really bullish on humans. And if you look at the kinds of things we do in our job... I mean The Jetsons was out over 50 years ago. We actually live closer to the time The Jetsons were set in than we live to when it was made. And they had Rosie-the-Robot, who went around and cleaned the windows and took care of little Elroy and did all of these things. And just think for a minute how far away we are from that. We don't even have robots that can have the coordination of a two-year-old, really.

And all of the complex things that people do, and as versatile as our brains are, I think people overestimate what technology is able to do, and, more to the point, vastly underestimate just the amazing versatility of humans. So yes, I think most jobs are robot-proof.

Tonya Hall: All right. There is a robot jobs test on your website. Tell us about it, and what you're actually learning from the results.

Byron Reese: So it's 10 questions, and it says, "You imagine a job." And then it says, "On a scale of one to 10, does this job require you to be in multiple places? Does it require emotional connectivity? Does it require to manage people? And then it scores that and tries to figure out the likelihood that that job is susceptible to automation. I mean, we've had thousands of people take it. Very few jobs actually are susceptible to automation within the time frame of people who currently have that job.

I mean, 50 years out, who knows. But what we're seeing so far, is that in the end, relatively little is susceptible to automation. And the real takeaway is that if there's a job that a machine could do. Just imagine a job a machine could do. If you make a person do that job, that is literally dehumanizing, because you just said there's nothing about that job that requires a person. A machine could do it. And so we really want to get people out of jobs that machines can do, because those are almost without exception dehumanizing jobs.

SEE: Hiring kit: Robotics engineer (Tech Pro Research)

Tonya Hall: "Is the computer of the future a thing or a being? Will it exist in the world, or will it experience the world?" That's a quote right from your book. So what's the answer?

Byron Reese: And the book asks you three philosophical questions that I can't answer for anyone, but you probably know the answers yourself, and then it tries to work through the implications. For that one, I would just pose one of them, and that is, "What are you?" And it's all multiple choice, so you get three choices. Are you a machine, and nothing happens in you that isn't just physics and electricity and chemistry? You're a controlled chemical reaction, and that's it. Or are you an animal, which is a mechanistic body with this thing, life, that we don't really understand? Or are you a human being?

And of course you're a human, but this means something more. You have a mechanistic body. You have life. And we have something else that makes us different. Some people think it's that we're conscious, that we experience the world. Some people think it's that we have a soul. Whatever you think. So if you believe we are machines, then a computer of the future would be a being. If you believe we're humans, the computer of the future isn't, because we can't be reduced to purely mechanistic terms. You cannot, in the end, be reduced to math in that case. So it all boils down to what you think the answer to that question is. What are you?

SEE: How to implement AI and machine learning (ZDNet special report) | Download the report as a PDF (TechRepublic)

Tonya Hall: So how do you define consciousness? Will a computer actually achieve that state?

Byron Reese: So consciousness... it's a myth we don't know what it is. Everybody agrees what it is. What it is, it's the experience of being you. It's the taste of a strawberry. It's the feeling of warmth. A computer can measure temperature, but the computer doesn't feel warmth. And so that is what it is. How it comes about, that's the question. In fact, it's arguably the greatest scientific question we don't even know how to ask yet, and we certainly don't know what the answer would look like. And so there are all kinds of theories. I go over eight of them in the book, and I try to, for each one of them, say whether a computer would be able to achieve consciousness if it is this.

Tonya Hall: Tell us about your most recent revelations regarding artificial intelligence. What's up and coming?

Byron Reese: Well, I would say that I have all of these people on my show, on "Voices in AI," and they're all practitioners in the field that are either in the university, or they're in business, or they're all people who are in the field. And there are two kinds of AI. There's one which is narrow AI--that's what we have everywhere today. That's your spam filter, and that's what routes you through traffic. But you never ask your spam filter what you should have for dinner, right? I mean, it's a very narrow thing, and we're really getting pretty good at that.

But then there's something else called general intelligence, and that's like a MacGyver sort of intelligence. It can do anything a human can do. And that's the one that some people are afraid of. But it's a very different thing, and so I always like to ask my guests if they believe we're on our way to building one. Because if you ask people when we'll get it, you get something between five and 500 years, depending on who you ask, which just kind of suggests that we may not know. Nobody knows how to build it.

But I think the interesting question is, are we even working on it yet? Are the techniques that we are using to build spam filters and route you through traffic is that eventually going to evolve into a general intelligence, or is a general intelligence something that is a completely different thing? And I would say my guests are evenly split on that. So not only do we not have a general intelligence, we may not have even begun to build one.

SEE: Robots with soft hands will transform the world. Here's why. (ZDNet)

We don't know how it is that people are intelligent. I mean we know we have a brain, right? But we don't know, for instance, how thoughts are encoded. We don't know how it is that we take knowledge from one area and apply it to another area so effortlessly. There is not even a consensus definition on what intelligence is. And so I think it's a pretty safe bet we'll never be able to engineer intelligence until we understand it better. And so I would say my revelation is I'm increasingly humbled by how little we understand even what intelligence is.

Tonya Hall: I really appreciate your time and your insights; deep thinking about the future of robots and jobs and artificial intelligence. And if somebody wants to connect with you, maybe they want to get a copy of your book, or maybe they want to follow your podcast, how can they go about doing that?

Byron Reese: Well, I'm the easiest person in the world to find. I'm Byron Reese at everything, so on Twitter, on Facebook. So just follow me there and visit, or go to

Also see