Erik Brynjolfsson Explains How Technology Drives Inequality And The Scary Truth About Robots Taking Human Jobs

Erik Brynjolfsson

This year at the World Economic Forum in Davos, participants have been largely focused on geopolitics, as economics takes a back seat.

But without a doubt, the No. 1 economic story that everyone is talking about — or at least the story that people say they’re talking about — is inequality.

MIT professors Erik Brynjolfsson and Andrew McAfee have just published their book “The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies,” which looks at the impact of technology on society, and discusses the role technology plays in work and inequality.

I sat down with Brynjolfsson to discuss his book, and dive into the rising impacts of digitization, robots, the winners and losers from all this, and how government policy ought to adapt to these changes.

One question that tends to cause a lot of anxiety has to do with which skills and industries will be accomplished by robots, and in which areas humans are safe. Brynjolfsson’s answer was a bit scary: Every time he thinks he’s found an area where robots can’t possibly compete with humans, someone in a lab at MIT or somewhere else is working on exactly that.

As for policies that government can embrace to avoid an extremely unequal distribution of the benefits of these technologies, Brynjolfsson brings up ideas in tax policy, education, and the promotion of entrepreneurship.

The full Q&A is below.


Screen Shot 2014 01 24 at 11.07.52 AM

What are you telling people about at Davos?

Well we’re telling them we’re in the midst of what Andrew McAfee and I are calling “the second machine age.” Just as the first industrial revolution automated and augmented muscle work, we’re now doing the same with mental and cognitive tasks. There are some similarities in terms of the potential growth and disruption, but there are also some big differences.

Who is getting hit hardest by the rise of robots?

The data say that routine information processing tasks have been especially hard hit. Think of bookkeepers or travel agents or tax preparers. That has been some of the first wave of how the “second machine age” has been affecting work, but what Andy and I talk about is just this whole wave of new technologies that is affecting all sorts of professional tasks like diagnosing diseases, legal briefs, affecting tasks that require mobility and fine motor control. We recently rode in Google’s driverless car, we visited Rodney Brooks’ “Rethink” factory where they are making Baxter robots that can do a lot of precision tasks. So there’s a whole set of categories that previously weren’t affected that now are being affected.

What are the skills that a human can have that won’t be replaced?

That’s a really tough question actually, because Andy and I have been playing that game for a couple of years, and every time we think we have a task that can’t be done, we find out shortly thereafter that there’s somebody in a lab at MIT or Silicon Valley that’s working on a prototype 0.1 of exactly a machine to do that task.

…every time we think we have a task that can’t be done, we find out shortly thereafter that there’s somebody in a lab at MIT or Silicon Valley that’s working on a prototype 0.1 of exactly a machine to do that task.

So it is a little scary.

That said, there are some areas that are certainly much much more difficult. Interpersonal relationships is something that machines just aren’t very good at. Humans have evolved for millions of years to be tuned to social cues. Creativity, coming up with a great opera or piece of art or piece of literature is something that machines can’t right do. Scientific discovery.

Also the nimbleness that the human body has still is not anywhere close to being matched by machines. But of course the technology is marching ahead and it’s accelerating so that the next 10 years are likely to have even bigger disruptions than the past 10 years.

Income inequality has grown and it’s a major issue. What percentage of the issue would you assign to technology?

Well there are a lot of forces affecting inequality. There’s globalization, there are institutional changes, cultural changes, but I think most economists would agree that the biggest chunk of it is due to technology. And that’s because of what economists call skill-biased technical change — favouring skilled workers versus less-skilled workers.

Also we talk in the book about capital-biased technical change — you bring capital over labour like when you replace humans with robots. And the third category that maybe is the most important one, we call it superstar-biased technical change, maybe we should come up with a better name. But it’s the fact that technologies can leverage and amplify the special talents, skill, or luck of the 1% or maybe even the 100th of 1% and replicate them across millions or billions of people. In those kinds of markets, you tend to have winner-take-all outcomes and a few people reap enormous benefits and all of us as consumers reap benefits as well, but there’s a lot less need for people of just average or above-average skills.

What long-term policy ideas should governments embrace to ensure that the benefits of new technology don’t just accrue to a narrow elite?

The reason Andy and I wrote the book is cause we want to change the conversation to get people focused on those kind of policy ideas. Cause the technology is racing ahead and it’s great what the technologies are doing in the sense that it’s making the pie a lot bigger. But policymakers are stagnating and people aren’t keeping up with their skills. We could do a lot more in reinventing education. We could do a lot moor in boosting entrepreneurship. We could do a lot more in changing our tax policy to favour people creating more jobs.

To touch on each of those a little bit more in depth, on education we need to cultivate those kind of creative skills that are more in more in demand. The word “job” really only existed for about 300 years and it really boiled down to making people almost cogs in a machine. Those kinds of routine instruction-following jobs are being automated away. What’s more important are creative skills and our schools aren’t structured to teach those.

When you think about it, for the 20th century they were focused on getting people to sit quietly in rows of desks and just follow instructions as best they could. That’s not the kind of education we need going forward. I think digitization can help with that.

On entrepreneurship, we’re seeing more and more entrepreneurs because the jobs are disappearing, but the real reason to try and boost entrepreneurship is not because everyone’s going to become an entrepreneur or should become an entrepreneur, it’s because ultimately those are the people in charge of inventing the new industries we need to create new jobs.

Just as Henry Ford helped create a new industry that employed millions of people who were previously working on farms and Steve Jobs, Bill Gates, others helped invent new industries, we need to invent the next new industries of the second machine age if we’re going to find work for all the people whose jobs can now be done by machines quite well.

The third category is tax policy and in that category, one of the basic laws of economics is that if you want less of something you tax it and if you want more of something you subsidise it. Right now, we tax labour for about 80% of our revenues one way or another in the United States. We put a wedge between what an employer has to pay and what an employee actually receives. And then we saddle them with a lot of required benefits. Each of these policies essentially discourages an entrepreneur from using labour to get the job done. Maybe that didn’t matter so much in the 19th and 20th century, but going forward, I think we need to flip that around. Remove the extra burdens we put on hiring labour and instead look at ways to at least make it neutral or better yet even subsidise it.

What surprises you in your research?

What surprised me most was how incompetent I was in understanding how rapidly technology would change. I study this as a job at MIT, yet even I have been surprised many times by how quickly the technology advances faster than expected.

10 years ago, I was teaching a class and part of it was could machines do well and what could humans do well. And I gave driving a car as an example.

10 years ago, I was teaching a class and part of it was could machines do well and what could humans do well. And I gave driving a car as an example.

You can’t just write a program for that. There’s no pre-set number of steps like playing chess. Lo and behold, two years ago, eight years later, I was riding down Route 101 in, yes, a driverless car. I hadn’t expected it to happen that fast. And frankly, none of my colleagues expected it to happen that fast.

Same thing with having machines you can talk to and they understand what you’re doing and carry out your instructions. Many of us carry those around in our pocket. Or having machines that can play Jeopardy, like IBM’s Watson, and win, and now also diagnose diseases and solve legal questions. So I’m trying to adjust to keep up with the pace of technology, and as I make those adjustments hopefully I won’t be as surprised much in the next 10 years cause now I have very high expectations.

Are you an optimist or a pessimist?

So the people I run into, they do tend to fall into those two camps. Some people are really excited about the future, especially technologists, the techno-optimists. I also run into a lot of economists. I just came from the American Economic Association, where I was on a panel with some economists who all had a very dismal view. In fairness, some of the economic statistics on median income and employment have been fairly dismal. Andy and I try to resist falling into either of those camps.

We call ourselves “mindful optimists,”

We call ourselves “mindful optimists”

in the sense that we think that the outcome can be very good. We can have a lot more wealth and have it be shared prosperity, but the mindful part is also important. We’re going to have to adjust our policies, our skills, our organisations to keep up with the technology. It’s not going to happen automatically. If we don’t start paying more attention to these issues, we won’t necessarily have that good outcome that we all hope for.