In the Age of AI

Artificial Intelligence - Inteligencia Artificial

In the age of artificial intelligence, predicting which jobs will fall to automation is as much about what machines can do as it is about what they can’t.

More than half of all jobs in America — both blue and white-collar — are resistant to automation, according to an acclaimed study published in 2013 by two Oxford University researchers.

Co-author Carl Benedikt Frey, who directs Oxford’s Technology and Employment program, broke down three areas where human intelligence still beats artificial intelligence: perception and manipulation, social intelligence; and creativity. Each type has what Frey calls a “bottleneck,” which slows the pace at which certain workforces can be automated.

The premise is simple: Technology won’t replace human workers if it can’t do the job. Six years later, Frey said the argument holds. Here’s a look at the sectors proving to be robot resistant:

Perception and Manipulation

Humans outpace robots when it comes to perception and motor skills, especially in so-called “unstructured work environments” – spaces cluttered with many different objects. Until technology catches up, work such as surgery is best left to humans, Frey said. Similarly, jobs in unstructured environments, such as homes, are more difficult to automate than jobs in predictable, structured environments such as factories, warehouses, airports and hospitals.

Social Intelligence

Social intelligence is another bottleneck, as machines can’t yet compete with humans at work involving negotiation, persuasion or care. In particular, machines struggle to recognize and respond to human emotions. Work requiring high social intelligence is therefore less susceptible to automation. Frey lists public relations and event planning as examples.


Robots also can’t keep up with human creativity: the ability to form new and valuable ideas such as poetry, music, recipes, jokes, fashion design or scientific theories. Though technology is capable of randomly combining old ideas to create new ones, the result doesn’t necessarily make sense — or have value. In this case, the novelty of a random, machine-generated idea should not be confused with creativity, Frey said.

What About the Rest?

Of course, that leaves roughly half the jobs in the U.S. with tasks machines can now do better than humans. But Frey emphasized that doesn’t mean those jobs will vanish outright, either.

“There are a lot of people that took our estimates to suggest that all of these jobs are going to disappear within 10 to 20 years, but the paper did not really say that,” he said.

If anything, technology will likely create new kinds of work for humans, says James Bessen, who heads the Technology and Policy Research Initiative at Boston University.

“People are focusing on the aspects of the technology where the machines can replace humans at certain tasks,” Bessen said. “Most of what technology does, is actually enhance humans at doing certain tasks.”

He uses bank tellers as an example. The profession was largely mechanized decades ago by the ATM – literally, “automated teller machine.” Yet, despite becoming a fixture at banks across the country, ATMs did not push out human tellers. On the contrary, research shows the number of bank tellers in America rose nearly in tandem with the number of machines.

Though ATMs did replace some workers, the cost savings allowed banks to open new branches for which they then hired additional people, Bessen explained. As such, the human jobs were not replaced but displaced. Moreover, the ATMs freed their flesh-and-blood colleagues from routine tasks such as depositing money. Instead, they were able to focus on inherently human work like interacting with clients.

Still, researchers acknowledge the transition to a new American workforce as these industries adapt to and integrate automation will be painful.

People will continue to lose jobs to new technology, and some workers may be pushed into long periods of unemployment, Bessen said. Already, many are forced to relocate or retrain, and those who can’t afford to do so will be left behind.

Women disproportionately hold the jobs at highest risk of automation, according to Molly Kinder, who studies the future of work at a D.C. think tank called Work, Workers and Technology.

“That’s not really being talked about,” Kinder said. “And that’s in part because women are overrepresented in some of these marginalized occupations like a cashier or a fast-food worker, and also in a large numbers in clerical jobs in offices.”

“There’s real pain involved,” Bessen said. “There’s a real question about how these gains get distributed, about who’s suffering, who’s bearing the brunt in these transitions.”

For American workers, finding ways to navigate the transition to new technology is the real challenge, Bessen said, adding that their success will in part be up to government policies — and pushback by workers themselves.