New technologies have been eliminating human jobs for centuries, but history shows that over time, automation brings more prosperity than hardship. At the start of the 19th century, for example, about 50% of the US workforce worked on farms. Today, the figure is just 2% -- a staggering loss of agricultural jobs to which the US economy, and its workers, have adapted quite well.
In part, this is because automation takes effect gradually. Steam engines ultimately powered the Industrial Revolution, but it took generations before people began using them to transformative effect. Even the humble telephone needed more than 50 years to work its way into half the households in America.
Admittedly, recent changes have been more sudden, with technologies such as driverless vehicles springing up seemingly overnight. And their economic impact will be dramatic. Drone freight trucks, for example, will eliminate the need for millions of drivers. But by driving close to 24 hours a day, driverless trucks could double the output of the US road transport network, while lowering its operating cost by 75% -- savings that will ultimately benefit consumers. By driving at a constant speed, these trucks will also burn less fuel, lowering carbon emissions. They will cause fewer accidents than human drivers. Are we willing to give up those gains in productivity, efficiency, environmental friendliness, and lives saved?
It's a serious question, as the effects on those who lose their jobs will be traumatic. But non-routine jobs are far less likely to be taken by robots. In the US, jobs involving knowledge work - managers, technicians, and professionals such as lawyers and accountants - have grown by nearly 2million per year.
And as always, while technology eliminates some jobs, it creates others, sometimes in surprising ways. One study, published in The Journal of Economic Perspectives, notes that, contrary to what might be expected, the number of bank clerks and bank branches in the US rose from 1980-2010, at the same time as ATMs were becoming universal. The reason: information technology spawned a host of new financial products, creating a need for new salespeople, administrators, customer relations staff, and of course, tech support. These new jobs more than offset the loss of the old ones.
What's more, many of the jobs likely to be eliminated are dangerous or unpleasant. Cleanup after disasters and bomb disposal can be handled by robots, saving humans from injury and death. Many industrial processes, such as manual welding, produce noise, heat, and toxic fumes, exposure to which can cause lung disease and kidney failure. Automating those jobs would improve the health of industrial workers. Automation also has the potential to expand the reach of vital social services where doctors, nurses, teachers, and home care professionals are in short supply. While most people might prefer to be cared for by a human, a robot nurse may be the best solution for those living in poor, rural, or otherwise underserved communities.
Automating tasks is not the same thing as eliminating entire occupations. Technology is already doing the former: McKinsey estimates that already-proven technologies could automate as much as 45% of the tasks individuals are currently paid to perform. Yet as the report notes, in many cases, automation will complement humans, rather than replacing them.
One example of humans and machines working together to solve complex problems is Centaur Chess, a system developed by chess Grandmaster Garry Kasparov after he lost a chess match to IBM's Big Blue computer in 1997. In this new game, a human chess player chooses from a list of moves suggested by a computer. The system has led to a higher level of chess-play than was previously possible by human or machine alone.
Chess is not a job, of course. But Centaur Chess shows that humans and machines working together on a challenging problem produced a better result than either the humans or the machines did alone. In the future, this type of collaborative analysis could lead to new forms of "augmented innovation" that solve complex problems and enhance productivity.
Even now, automation can serve as a complement to human labor. In the US, Germany, and South Korea, robot use and human employment have risen in tandem. And productivity gains delivered by automation and machine intelligence can benefit workers even more directly: one US company that installed robots to perform repetitive manufacturing tasks saw a 20% increase in productivity, allowing it to hire more employees.
In addition, robots must be built, programmed, maintained, repaired, and supervised, creating a whole new class of technicians and managers. The World Economic Forum projects a sharp rise in the need for specialized sales representatives, as technology becomes so embedded in every industry that companies will need a sales force capable of explaining the benefits of their own platforms across different industry sectors. And for each of these new talent categories, recruiters will be needed who are sufficiently well versed in the specialized technology of each niche as to recognize good talent when they see it, and to win over the most desirable candidates.
Given the new jobs that will be created, and the new skills that will be required to do them, many employees in today's workforce will need to adapt, if they want to stay employable. AT&T estimates that 280,000 of its own employees must be retrained - or, more accurately, must retrain themselves to learn new skills. CEO Randall Stephenson has advised employees to spend five to 10 hours a week in online learning, and AT&T reimburses employees up to $8,000 a year in tuition costs, defraying the costs of both short courses as well as full-blown graduate degrees. Courses popular with managers eager to hone their technical skills include data analysis, introduction to coding, and smartphone app development. AT&T promotes these courses aggressively via emails and video broadcasts, and has even provided partial funding for an online masters in computer science offered by Udacity, an online learning site, and the Georgia Institute of Technology.
In addition to private companies, schools will need to adopt a stronger vocational tilt when educating young people. Alec Ross, a former adviser on innovation policy to Hillary Clinton, argues in a new book that the US should scrap the vocational training programs which have prevailed in its community colleges for the last half century, dropping courses like carpentry and welding in favor of cyber security and IT system administration.
As robots go from performing work that is manual and repetitive to doing work that is cognitive and non-routine, people will have to be taught to think in ways that set them apart from machines. As the Institute for the Future says in its report, Future Work Skills 2020, we will have to ask ourselves fundamental questions such as, "What are humans uniquely good at? What are our comparative advantages? And what is our place alongside these machines?"
The answer to these questions may be found in the different types of intelligence uncovered by scientific research. Analytic intelligence uses logic, math, and algorithms - areas where machines hold an uncontested advantage. Creative intelligence is the ability to cope with new situations using existing knowledge. Contextual intelligence involves understanding the relationships between things in the world. Schools ought therefore to move away from curricula based largely on rote memorization (where computers and search engines excel well) to courses that cultivate contextual and creative intelligence.
Industry 4.0, where industrial manufacturing processes are connected to the Internet, represents the next stage of automation, one where smart systems operate under human control. But, along with the Internet of Things, it is unlikely to arrive in a sudden technological explosion. More likely, these new platforms will unfold in stages, giving us time to prepare.
Yes, tasks will be eliminated and jobs lost. But by augmenting and amplifying human capabilities, automation will make many people far more valuable than they could ever hope to be without robotic help. When human and machine capabilities are combined intelligently, productivity and living standards will rise, economies will grow, new industries and business models will emerge. All of those changes will improve the quality of people's lives. The benefits to consumers and businesses will be substantial.
While it's natural to have concerns about change, there is no compelling reason to stuff the technological genie back into the bottle. At the same time, people should cultivate skills and traits that (for now, at least) are uniquely human: creativity, empathy, the ability to draw inferences or connect disparate ideas.
Google's chief economist, Hal Varian, advises job-seekers to become "an indispensable complement to something that's getting cheap and plentiful." Automation will be both. But there is every reason to believe that human creativity, reason, and empathy will provide indispensable complements to the efficiency and tirelessness of our robotic colleagues.