As machine learning advances at exponential rates, many highly skilled jobs once considered the exclusive domain of humans are increasingly being carried out by computers. Whether that’s good or bad depends on whom you talk to. Technologists and economists tend to split into two camps, the technologists believing that innovation will cure all ills, the economists fretting that productivity gains will further divide the haves from the have-nots.
To set the context for the debate and to examine potential policy implications for pressing social questions, McKinsey’s Rik Kirkland conducted a series of interviews in January at the World Economic Forum’s annual meeting in Davos. Among those interviewed were Nobel Prize–winning economist Robert Shiller, author of Finance and the Good Society (Princeton University Press, 2012), among other books; data scientist Jeremy Howard; and professor Erik Brynjolfsson, of the Massachusetts Institute of Technology, who, along with Andrew McAfee, is the author of The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (W.W. Norton & Company, 2014). This edited transcript captures and combines highlights from those conversations.
Workforce at risk?
The Quarterly: Erik, your new book focuses in part on the social and policy implications of what you’re calling the Second Machine Age. Can you set the context for these issues?
Erik Brynjolfsson: In our book, we call it the great paradox of the Second Machine Age. The paradox is that even though we’ve had record levels of wealth creation, employment hasn’t kept up. Even though we have record productivity, US median incomes are lower now than they were in the 1990s. There’s been a decoupling of the two. As we researched the book, we found two disparate groups that tend to focus on either productivity or inequality.
On the one hand, there are the techno-optimists, who often have a utopian view of the world. This group thinks that technology will solve all our problems. In fairness, innovation has never been faster, and technology has solved an amazing set of problems. The conclusion this group sometimes draws from that is “just let technology do its thing and the rest will take care of itself.” People in the second group, which is largely composed of economists, tend to have a pessimistic view of both the present and, perhaps more disturbingly, the future. They point out that median income is lower now than it was in the 1990s, that unemployment has been struggling, that the economy, by many metrics, has been performing quite poorly. They believe that we’ll have less economic growth in the future than we did in the past.
As you try to reconcile those two groups—the utopians versus the dystopians, or “stagnationists”—you can end up concluding that, in many ways, they are both right, about the facts, at least, if not the conclusions. We do have innovation, but that doesn’t mean everyone’s necessarily going to be taken care of. We have stagnating median income, but that doesn’t mean innovation has slowed down.
In fact, the dirty secret of economics is that it’s possible for technology to make the pie bigger—and that’s exactly what it’s been doing. Record wealth. But at the same time there’s no law that everybody’s going to benefit from technology. Some people, even a majority, may be worse off. Ever since the Industrial Revolution, we’ve experienced a rising tide that has helped most people. But in the past 15 or 20 years or so, those trends have diverged. We have what my coauthor, Andy McAfee, and I call the Great Decoupling. Productivity has continued to grow steeply and innovation has been strong, but median income and employment have stagnated.
The Quarterly: Jeremy, how big a deal are machine-learning algorithms for employment and the workforce? And what should we do about it?
Jeremy Howard: I think it is important to think about the policy implications here. Government leaders need to be aware that, right now, computers are as good as or better than humans at most of the tasks people involved in information-processing jobs do. That is 65 percent of the American workforce. So is this wonderful or is this a tragedy? It actually depends entirely on how governments respond. Scenario number one is a disparity in economic power, in which the folks with the data and the algorithms have—and add all of—the economic value, and the rest of the workforce adds little or none.
That scenario could create an awful social disruption. Scenario number two is to accept that in this new world, there’s a large group of people who can’t really add economic value anymore, but that doesn’t mean they don’t get to live a decent human life. So we have to start thinking about the policy implications—like a basic living wage, which Germany will be introducing, or a negative income tax, which has been off the agenda for decades but deserves to be back on it. I think people should start to think about these policy implications because the point at which we need to make decisions will be upon us suddenly.
The Quarterly: Bob Shiller, you’ve devoted research time to the implications of technology for peoples’ livelihoods, and have discussed them in your recent book. How do you see the problem?
Robert Shiller: I think it’s the most important problem facing the world today. It’s associated with income inequality, but it may be more than that. Since we tend to define ourselves by our intellectual talents, it’s also a question of personal identity. Who am I? Intellectual talents are being replaced by computers. That’s a frightening thing for most people. It’s an issue with deep philosophical implications.
Addressing the issue: The view from New Haven
The Quarterly: What are we doing about the problem?
Robert Shiller: Politicians generally haven’t confronted it. Why? Well, our political system encourages them to pick points that get them elected. Furthermore, it isn’t easy finding narratives that will appeal to voters right now. Any story about inequality tends to be divisive, and as a politician you’d rather go with the fiction that we’re all going to be better off rather than that there will be inequality. Furthermore, coming back to my point about personal identity, any discussion around policies that deal with inequality threatens the sense of identity people have through their role in society. I think politicians have difficulty with this.
The Quarterly: So how can we get the ball rolling?
Robert Shiller: We need to start planning for how we will deal with the inequalities produced by advanced technology. We need to plan now and to enact legislation for what we will do about this in the future.
The idea would be to write legislation now that would be open ended. At any time in the future, when inequality rises beyond some threshold, then the tax system would automatically kick in. It’s automatic unless the legislation is rescinded.
Other countries have equalization programs in their law. Canada, in 1982, put into its constitution an equalization program requiring all the provinces of Canada to have substantially equal social services for their people. You don’t want the poor provinces suffering. And that equalization program is still in force now, more than 30 years later. Once it’s in the constitution, it has a moral force that is hard to change. So we need something like that, not that I think these equalization programs various countries have are strong enough. They tend to look only at government services and not at the whole range of problems that inequality might produce.
More specifically, one plan might be what I call inequality indexation of the tax system. This would automatically raise taxes on the rich in the future if inequality becomes much worse. Maybe that wouldn’t be so hard to put in place today, and it might be of great value in the precedent it sets for the future.
The Quarterly: How would that work? Are you referring to some sort of wage insurance?
Robert Shiller: Insurance is a fundamentally important concept. Whether it’s provided by the private sector or the government, it’s an organized response to uncertainty that affects individuals. Wage insurance is one form that has been experimented with by the US government, in fact, in an effort to deal with the problem that globalization is taking jobs out of the country. But it hasn’t been experimented with very much yet. I think we could have a much more comprehensive system of wage insurance. And we can do it without moral hazard if we manage it right.
The moral hazard is that people will stop improving themselves thinking they can collect insurance rather than work hard. But there are ways to design a program to minimize this.
But wage insurance is not exactly the right name for this. I call it livelihood insurance. By using that name we’re not talking just about wage earners, we’re talking about the earning potential an individual has. It may or may not be wage income.
It also, to me, helps convey the lifetime nature of the problem. People go through a life cycle. Early in the life cycle they train themselves into some particular occupation. Later in life, they very rarely undertake retraining. And they rise and fall with that occupation.
What I thought the most important to do would be to develop measures of occupational risk and to write insurance policies against risks to the occupation, and that eliminates moral hazard because a person can’t do anything about the rise and fall of the occupation.
So in order to make that happen we have to develop a rather different kind of insurance industry. This has never been done, and it requires better measurement of occupational incomes. We have to measure risks before we can insure them. And we need to develop historical indexes of occupational incomes that are done properly so that we have some sense of history about how these things evolve. That will help us design insurance policies.
It’s going to be much harder to deal with these problems if we wait until 70 percent of the population is out of a job. Then it might be very difficult to get a consensus on what to do about the problem. This is really the insurance principle. You want to insure a house against fire before it catches fire. After it catches fire, it’s too late.
The Quarterly: What would you like to see happen from an economic-policy standpoint to foster job creation?
Robert Shiller: The US is one of the entrepreneurial centers of the world, but even so it may not encourage entrepreneurs enough. I’m thinking particularly of financial innovation. I had the experience of trying to launch a new financial product and then dealing with US regulators, whom I found sympathetic because the product was socially minded and didn’t come across as an attempt to make money, particularly. But I found that the regulatory system is kind of bureaucratic and slow to respond. The regulators wanted to fit us into one of their existing structures. That seemed very difficult, and it involved long delays to come up with anything really new. Now, this is a fundamental problem, I guess, because financial innovation has at various times in history been exploitative, so there’s a legitimate reason to be concerned about that. I think dealing with this takes resources. One thing we have to recognize is that government should expend resources on its management of innovation.
Another example is the patent office, which needs to have capable people with time to really study patents. The problem has been that the government is not willing to allocate enough funds to the patent office to make its activities as enlightened as they should be.
Regulators also are under pressure. Former Securities and Exchange Commission chairman Arthur Levitt wrote a book called Take on the Street1 detailing the veiled threats he got from financial-interest groups that didn’t want to be regulated. People were nasty at times. And they would try to sneak into legislation and regulations little words, which nobody else knew the significance of, that would weaken the legislation or the regulations.
So I tell my students that well-meaning people who want to improve society might consider careers as regulators. I think that our society undervalues them—that they’re very important. At least it’s like joining the Peace Corps. I think young finance people ought to spend a few years working at a financial regulator.