James ManyikaMichael ChuiKaty George, Feb 03, 2017
Soon after taking office, the new president created a national commission to examine the impact of automation. No family should pay an unjust price for progress, he announced, yet automation should not be viewed as an enemy. “If we understand it, if we plan for it, if we apply it well, automation will not be a job destroyer or a family displaced. Instead, it can remove dullness from the work of man and provide him with more than man has ever had before.”
The U.S. president who spoke those words was Lyndon B. Johnson, and the year was 1964.
A half-century later, technology has advanced at breakneck speed. Who back then, other than science fiction writers, could have imagined Amazon’s drone shipments, the legions of robots at work today in manufacturing, or the algorithms now being used to detect cancers? Yet anxiety about automation is still with us. Today there is concerned debate about the impact of technology on the economy and especially on the future of work.
It’s instructive to note how the economy has continued to prosper, and people have continued to work, since the 1960s, even as the workplace itself has been reshaped by technology. New jobs that could not have been imagined at the time, such as app developer or MRI technician, have replaced obsolete ones, such as switchboard operators. That’s a pattern we have seen since the beginning of the Industrial Revolution, two centuries ago, when more than 60% of Americans worked on the land; today it’s less than 2%. Still, we cannot help but wonder: Could this time be different?
We have just published new research about automation’s potential effects, based on an in-depth analysis of more than 2,000 workplace activities across 800 occupations. We focused on activities because all occupations consist of numerous activities, each of which can be automated to a varying degree. Within marketing, for example, some tasks can be automated easily, but others cannot.
We found that half of the activities people are paid to do in the global economy have the potential to be automated using current technology. The most automatable activities involve data collection, data processing, and physical work in predictable environments like factories, which make up 51% of employment activities (not jobs) and $2.7 trillion of wages in the U.S. These activities are most prevalent in sectors such as manufacturing, food services, transportation and warehousing, and retail.
More occupations will change than will be automated in the short to medium term. Only a small proportion of all occupations (about 5%) can be entirely automated using these demonstrated technologies over the coming decade, though the proportion is likely to be higher in middle-skill job categories. But we found that about 30% of the activities in 60% of all occupations could be automated — and that will affect everyone from welders to landscape gardeners to mortgage brokers to CEOs. We estimate that about 25% of CEOs’ time is currently spent on activities that machines could do, such as analyzing reports and data to inform decisions.
Automation’s potential is broader than it historically has been because technologies including robotics, artificial intelligence, and machine learning are increasingly able to accomplish not just physical activities but also ones that include cognitive capabilities, from lip reading to driving. As companies deploy automation, we need to think more about mass redeployment than unemployment, and we need to equip people with the skills they will need for the workforce of tomorrow — including being able to interact much more closely with machines in the workplace. These machines have capabilities that are inherently human, including managing and developing people, along with social and emotional reasoning.
Like President Johnson in the 1960s, we see that automation could make a major contribution to productivity and prosperity. Our research suggests that future automation could raise productivity growth globally from 0.8%–1.4% annually, which can make a meaningful contribution to global economic growth and compensate for the demographic headwinds of aging populations. For companies around the world, automation will offer the potential to capture substantial value — and not just from labor substitution. These technologies enable higher throughput, enhanced quality, better outcomes, greater safety, and the opportunity to scale up or adopt new business models.
However, just because the technological potential to automate a workplace activity exists does not mean that it will happen anytime soon. The pace and extent of automation will depend on a range of factors, of which technical feasibility is only one; there are still important barriers to overcome, including the ability of computers to generate and understand natural language. Other factors include the dynamics of labor supply and demand. If there is no shortage in the labor market for lower-wage cooks, for example, it may not make business sense to replace them with an expensive machine.
The benefits for business are relatively clear: better, smarter, error-free outcomes, along with innovation, productivity, and growth. For policy makers, the issues are more complicated. They should embrace the opportunity for their economies to benefit from automation’s productivity growth potential, and put in place policies and incentives to encourage investment in continued progress and innovation. At the same time, they must enact policies that help workers and institutions adapt to the changes in employment. This will likely include rethinking education and training, income support and safety nets, and transition support for those dislocated. Above all, a focus on the skills needed to thrive in this new era will be paramount. The lesson from history is that innovation, investment, and growth create demand and jobs that may once have seemed like science fiction.