MIT’s Erik Brynjolfsson on the automated future of work.
Automation is a new mass economic problem, one that arguably affected the outcome of the presidential election. Thousands of Americans have already been displaced by a force of much more efficient workers—machines. As the full impact of digital technologies is felt, jobs of all kinds, from real-estate agent to pilot to truck driver, will be upended.
Erik Brynjolfsson is the director of the MIT Initiative on the Digital Economy, a professor at the university’s Sloan School of Management, and a research associate at the National Bureau of Economic Research. He was one of the first researchers to study IT productivity and quantify the value of the variety of products online (aka the “long tail”). Brynjolfsson’s 2014 book The Second Machine Age (co-written with Andrew McAfee) looked at how digital technology is transforming our jobs and our lives. The authors argued that automation is advancing to the detriment of today’s economy, yet for the potential good of the environment.
CityLab recently spoke to Brynjolfsson about artificial intelligence, the robotic future, and the possibility of a “leisure society”—one so efficient at production that citizens don’t need to work.
Machines are replacing manufacturing workers. The old factory jobs are gone, and most economists say that they are never coming back. What does this mean for the way we organize society?
Most manufacturing jobs are never coming back. America manufacturing output has never been higher, and the decline in jobs does not represent the work going overseas, but rather the increased automation and productivity of American factories. Just as agriculture accounted for 42 percent of the workforce in 1900 and less than 2 percent today, manufacturing labor is becoming more productive, and so fewer factory workers are needed.
If we want to put people to work, the answer is not to try to restore the jobs of the 20th century, but to invest in creating the new jobs of the 21st century, which are increasingly in the service sector, healthcare, education, and creative fields. We shouldn’t try to protect the past from the future by protecting incumbent companies from new technologies and new entrants. Instead, we should be investing for a rapidly emerging future with confidence and high ambition.
Do you believe that we’re going to have a “leisure society”? Or is there an alternative path?
We will eventually have a leisure society, assuming we don’t blow ourselves up first. But we must remember that we are still far from a world without work—there is no shortage of work that needs to be done on our planet today, from taking care of children, the elderly, and the sick, to cleaning up the environment, to making scientific and medical breakthroughs. Much of this work can only be done by humans, at least with current technologies and those likely to be available in the next few decades. So I sometimes worry when people jump too far ahead and propose policies for a world where we don’t yet live.
Elon Musk has said that a universal basic income may be inevitable. He predicts robots will be the bulk of the workforce of the future, so the U.S. government will need to pay non-working humans a livable wage. How probable is that?
I agree with Elon Musk that we will almost surely have some sort of universal basic income when machines reach the point where they can do most or all of the work that humans do. But that’s not where we are today. And we likely won’t be there for decades.
Because there’s so much work still to be done which only humans can do, and because people get a great deal of meaning and satisfaction from being contributing members of society, I prefer an approach like the Earned Income Tax Credit, which rewards people for staying engaged in the workforce and rewards employers for finding and organizing useful work for them to do. The EITC is a great way to reduce inequality while keeping millions of people engaged in the economy.
In 50 years, which of today’s jobs will not be doable by machines? What will be the path to a meaningful life in the imminent robot economy?
In 50 years, there may be very few jobs which can’t be done by machines. It will likely be a world with vastly more wealth [because technology will make it possible to produce goods and services in abundance at low cost] and vastly less need for work. Shame on us if we somehow turn that into a bad thing. I believe it can, and should, be the best thing that ever happened, giving us the resources to address problems of disease, poverty, environmental degradation, and scientific puzzles that humanity has struggled with for millennia.
We can have more leisure and more time for interacting with friends and family, playing games and sports, reading, and thinking. What’s more, smart machines can help us have a lighter footprint on the planet, whether by reducing energy use through greater efficiency and lower cooling bills, or by replacing atoms with bits for media, entertainment, travel, and other goods and services. This good outcome is not inevitable, but will depend on us adapting our economy, the way we create and distribute income, and the way we get meaning from life.
Do you believe that we can leave the resolution of the problem to natural economic forces or do we need to intervene? Is there a new doctrine you can identify that will better serve the American people?
The idea of “natural” economic forces is misleading. The market system depends extensively on a variety of formal and informal institutions, norms, laws, practices, and decisions by millions of people.
When the first industrial revolution augmented muscles with machinery, great wealth was created, but there was also an enormous amount of disruption. Millions of jobs were lost and millions more were created, as people moved from agriculture to industry. American society responded with a variety of changes, including the introduction of mass public education, first via primary schools and later high schools.
These ideas were considered radical at the time. They were essential to provide the literacy and numeracy required for an increasingly industrial economy, and created not only prosperity, but broadly shared prosperity. Later changes in antitrust laws, a national income tax, social security, various labor laws, and many other adjustments continued the adjustment process needed take full advantage of these new technologies.
[Today’s] technologies are augmenting human minds, not just muscles, and I expect equal or larger effects on the economy and work. As in the past, our institutions will need to adapt. In particular, we need to reinvent education to foster creativity as well as interpersonal skills like motivation, persuasion, teamwork, caring, and leadership. We need to boost entrepreneurship so that new goods, services, and industries can be created as jobs in the old ones are increasingly automated. And we need to rethink our tax and transfer systems so that all members of society have an opportunity to benefit, not just a few.
In your writing, you seem to see increasing automation as an opportunity, not a disaster. You argue that it’s perhaps a positive thing, ultimately, that robots are taking our jobs. This seems counterintuitive, but is there reason to be optimistic?
This is not the time for unconditional optimism. Nor is it the time for unconditional pessimism. We must always bear in mind that technology does not decide how we create or allocate wealth. Instead, technology is a tool.
If we decide to use the power to create shared prosperity, I’m confident we can do so. But it will only happen if we actively work towards that goal. As one small step towards [it], we’ve created the Inclusive Innovation Challenge at MIT to recognize and reward individuals and organizations that are using technology to benefit the many, not just the few. This is in keeping with the principles we outline in our open letter on the digital economy.
We have more powerful tools now than ever, which means we have more power to shape the world.