Are social skills the last line for human job-hunters?

Photo / iStock

Dennis Mortensen, the founder of a technology startup, had a simple request for Amy Ingram: Set up a meeting with a journalist.

Within minutes, Amy sent an e-mail with a preferred date and time and two alternatives. She reached out again in the middle of the night about nine hours later after getting no response: "I wanted to follow up with you about this meeting with Dennis."

Four hours after that, she fired off another e- mail: "I haven't heard back from you yet about this meeting. Is this time convenient?"

Some would call Amy persistent, others might say she's overly aggressive or even annoying. What can't be argued is that she got the job done. A meeting was scheduled less than 24 hours after the initial request.

Amy Ingram is a virtual personal assistant created by New York-based Mortensen and his team at x.ai. Her initials are also those of artificial intelligence, and her last name refers to a model used in helping machines understand human speech.

Her single-mindedness is an example of the need to imbue such programs with social skills, so Amy remains a work in progress, Mortensen said in the interview arranged by her.

Efforts to make agents of artificial intelligence more sentient threaten to erode one of the biggest competitive advantages humans have over robots: the ability to summon social skills, including reading a person's feelings, managing emotions, working in teams and communicating effectively, to complete a task.

Machines' capacity to do routine jobs that don't require a personal touch - think assembly-line workers, bank tellers or grocery store checkout clerks - has been behind the so-called labor-market polarisation that has contributed to the hollowing out of middle-class professions.

David Deming, an associate professor of education and economics at Harvard University, has found that jobs requiring social interaction are growing relative to work that doesn't, and such skills may offer some protection from robotic takeover.

Certain high-level professions that demand technical expertise and low-skill work that can be done by a greater share of the population often have in common a need for language, creativity, flexibility and physical dexterity, all things humans currently can do better than machines.

Almost all job growth since 1980 has been in work that is social-skill intensive, according to research from Deming published in August. Occupations that require high levels of analytical and mathematical reasoning but little social interface - for example statistical clerks and machinists - have "fared especially poorly," he wrote.

Meanwhile real wage growth has been strongest in jobs that require workers to have both math and social skills, such as registered nurses, designers and financial managers.

One day that computer on the other end of the customer service line will be so good you won't need to keep pressing zero to reach a human operator.

Social interaction is "an unconscious process" for people, Deming said in an interview. "It's really hard to write a program that does that as well."

While true for now, it's only a matter of time before robots catch up to humans in this area too, argues Pedro Domingos at the University of Washington in Seattle. He says machines are already making impressive headway on at least mimicking social skills. One day that computer on the other end of the customer service line will be so good you won't need to keep pressing zero to reach a human operator.

In Domingos's labor market of the future, having a person do something a robot can do for less -- tend bar, wait tables -- will be a luxury. Jobs like that "will remain, but they're going to be comparatively very highly paid jobs and there are going to be fewer," he said.

Even technical workers such as computer scientists will be out of work eventually, as machines become more nimble at understanding natural language, said Domingos, who is the author of "The Master Algorithm," a book on machine learning published last month.

For David Autor, an economics professor at Massachusetts Institute of Technology in Cambridge who is considered a leading expert in the field of job polarisation, the future is brighter, although there is likely to be some hardship in the form of lost jobs along the way. Automation has put people out of some type of work - agriculture, manufacturing, dishwashing - for the past 200 years, and mankind always bounces back, he said.

"People tend to underestimate that, as we mechanise one set of things, we think of all kinds of new things to do," Autor said in an interview. The way we respond "is through our creativity and also through educating ourselves. We continue to make ourselves relevant."

People tend to underestimate that, as we mechanise one set of things, we think of all kinds of new things to do.

Pedro Domingos

In the meantime, women may have the upper hand. Jobs that require social skills have become more female-dominated, and women consistently score higher on tests of emotional intelligence, according to Anita Williams Woolley, an associate professor of organisational behaviour and theory at Carnegie Mellon Tepper School of Business in Pittsburgh.

That may help explain why labour force participation for men in their prime working years is hovering at an unprecedented low near 88 percent, she said.

"To some degree, you should see women start to have an advantage," Woolley said.

Mortensen agrees, though his invention may soon put some out of work. In 2014, about 94 percent of secretaries and administrative assistants were women, according to Labor Department data.

Asked how he feels about Amy being a potential job-killer, Mortensen was quick to say that few people are lucky enough to have an assistant - he's just trying to democratise the job.

The hope is that humans will no longer have to spend time on "e-mail ping-pong," Mortensen said. "I still think there's room for us, and at the next level. That's just the optimist in me."