Skip to main content
the future of work

In his famous laws of robotics, science-fiction writer Isaac Asimov decreed that robots cannot hurt humans, but he didn't say anything about not taking our jobs. Automation in the workplace isn't new – economist John Maynard Keynes warned of "technological unemployment" as far back as 1930, and even that came more than a century after mechanical looms riled up the first Luddites – but recent advances in artificial intelligence have allowed robots to climb the corporate ladder. How high can they go? Is there a corner office in R2-D2's future?

"I believe we're going through the same process we went through in agriculture, but for knowledge work and information work," says Marina Gorbis, the executive director of the Institute for the Future, a non-profit forecasting and research organization based in Palo Alto, Calif. "A lot of things that were routine are becoming automated. We're seeing this with automated sales calls, and we're seeing it in offices, where administrative work – like scheduling meetings, and other kinds of co-ordination – can be done with software."

"We used to think of automation as just affecting those lower level kinds of white-collar work," adds Andrew McAfee, principal research scientist at the Center for Digital Business in the MIT Sloan School of Management in Cambridge, Mass. He's the author, with MIT colleague Erik Brynjolfsson, of Race Against the Machine , a study of how the digital revolution is transforming employment. "If all you were doing was making sure all the shipping invoices got lined up, that job got automated away a while back. Now, what we're seeing is the same thing happening to higher-end, higher-education, higher-wage white-collar jobs."

Dr. McAfee points to the rise of income tax preparation software. Law firms, too, are feeling the artificial intelligence pinch: As pattern recognition text-reading software improves in accuracy, Dr. McAfee says, we're seeing an end to the days of "huge teams of discovery lawyers sitting in a gym, reading through boxes of documents looking for keywords."

But there may be a limit. Frank Levy, a lecturer in the Department of Health Care Policy at Harvard Medical School in Boston, sees computers as often complementing, rather than substituting for, human skills. "In most cases, computerization seems to take a piece of a job," says Dr. Levy, who co-authored with Richard Murnane The New Division of Labor, a comparative study of human and computer workers. "Computerization often takes a task, so it takes jobs indirectly. For example, the ATM [automated teller machine] took one part of a teller's job and so banks needed fewer tellers."

Dr. McAfee studies how quickly we can (or should) turn such tasks over to machines. He doesn't see management-level jobs making that jump.

"When I think of managers, I think of the classic work of bringing a group of people together, making sure they're working well together, getting the most out of them and directing them toward a goal," he says. "Mid-level managers have been a fairly durable group when you look at how many of them you need to do a million dollar's worth of work. I've never seen a computer that is good at that. Yet."

But just because those managers are sticking around the office, it doesn't mean that automation won't be changing their work lives. As technology takes on a larger role in the office, management will face new challenges – not from the computers that have taken over jobs, but from the workers left behind.

"It slows an economy to hold onto tedious jobs that are better done by machines," says Michael Leiter, a psychology professor at Acadia University in Wolfville, N.S., the Canada Research Chair in Occupational Health and Well-Being, and an international consultant specializing in increasing employee engagement.

Refusing to offload tedious tasks onto technology, he says, "interferes with employees' capacity to fulfill their potential and to keep on the cutting edge of their profession. I know first-year law associates for whom digging through e-mails and documents is the bane of their existence."

Giving grunt work to the robots can make for happier and more productive employees (assuming they weren't laid off) but only if it's done right. To see that boost in engagement, and to benefit from it, managers will need to develop "new areas of activities that allow those workers to function on higher levels of sophistication," Dr. Leiter says.

As managers strive to keep workers engaged in an increasingly robotic workplace, Dr. McAfee at MIT wonders whether their success may hinge on the one thing computers don't have: their human touch.

He is quick to clarify that he doesn't buy the argument that "human intuition is this ineffable special thing that we need to preserve. Human intuition, as far as we know, is pattern matching. Full stop. We're good at it, but the triumph of [IBM supercomputer] Watson winning Jeopardy! convinces me that computers will be very good pattern matchers." And, when possible, he sees an inarguable benefit to eliminating human frailties such as sleep deprivation, personal problems and sickness: "The more I understand human biases, the more I wonder why we're letting people make those decisions that a computer is consistent at making."

Still, when it comes to managers motivating and engaging their work force, there may be an irreplaceable value to old-fashioned human-on-human interactions.

Dr. McAfee sees parallels between the manager-worker relationship and the one between doctor and patients. "What a doctor does when he diagnoses me is listen to my symptoms, pattern match that against all the stuff that could be wrong with me, and come up with an answer," he says.

"Computers are very good at that. But for many people, the process of talking to a human doctor is an important part of getting them to comply with the treatment." They talked to someone who was an authority figure, who listened to them, who told them, 'Take two of these pills for the next five days.' It's that human interaction that makes them more likely to actually take two of those pills for the next five days."

"I just read an article that said IBM's Watson can diagnose cancer better than a doctor," adds Ms. Gorbis, of the Institute for the Future. "But we're physical beings, and there are certain things we can only get from human interactions."

Interact with The Globe