Amid Automation Debate, A.I. Backers Tout Job Creation Potential
Rapid advancements in robotics and artificial intelligence technologies over the past decade have stoked concern that machines could eventually take over most, if not virtually all, human jobs.
But there is another, more optimistic view of how the advance of A.I. and automation will impact the economy—one articulated by people like Mark Gorenberg (pictured right). He’s a managing director of Zetta Venture Partners, a San Francisco-based firm that invests in “intelligent enterprise software” startups.
He argues that as A.I.-related technologies advance, it’ll be increasingly common for machines to work side by side with humans and “empower” them to perform at a higher level. Technology advances will likely result in fewer jobs in some areas and more jobs in others, he says. “Our take is, net, it’s going to be positive for society and positive for the human condition,” he says.
“I don’t think we should fear these technologies,” Gorenberg continues. “I don’t think A.I. is going to be the chief threat that people fear that just eliminates jobs.”
A.I. has become one of the most active technology sectors in the past few years, with venture capitalists and large companies investing heavily in the field. Amid the hype, Gorenberg’s comments are worth highlighting because they provide a window into how some investors view A.I. and its potential impact on businesses and society.
Of course, no one can predict with certainty how things will play out. As a Zetta spokesman points out, there are wildly differing estimates for how A.I. technologies might affect human jobs. A paper published last year by the Organisation for Economic Cooperation and Development estimated that an average of 9 percent of jobs across 21 countries could be susceptible to automation. Meanwhile, a University of Oxford study from 2013 estimated that 47 percent of U.S. jobs are at risk of being taken over by computers.
Erik Brynjolfsson, the director of the MIT Initiative on the Digital Economy, thinks that machines could eventually render human workers obsolete in virtually every industry, but that shift will take at least 30 to 50 years to play out. James Bessen, an economist and lecturer at Boston University’s law school, has written that computers automating tasks doesn’t necessarily imply significant job losses. The bigger and more immediate challenge, he has argued, is making sure humans learn the skills necessary to use new technologies.
Gorenberg, a member of MIT’s board of trustees, says he doesn’t know if A.I. technologies will lead to more or fewer jobs overall. “But I don’t subscribe to the camp that says it is inevitable that it will lead to more, or much more, net unemployment,” he says.
Some manufacturing and warehouse work will continue to be automated; self-driving cars and trucks could eliminate the need for many human drivers; and cashiers will likely continue to be replaced by self-service kiosks, Gorenberg says.
But in other occupations, Gorenberg thinks intelligent software and machines will augment the capabilities of human workers and, in some cases, actually help create jobs for people.
In healthcare, more sophisticated technology could enable a larger pool of candidates to attain a job in the industry because the machines will be able to handle tasks that fill in gaps in people’s skills or experience level. One example might be enabling more people to operate ultrasound equipment, he says. He acknowledges that it “may be possible” that such machines might eventually become sophisticated enough to run themselves without human oversight, but that isn’t going to happen any time soon.
Gorenberg gives another example in sales software. His firm has invested in InsideSales.com, a Utah-based company whose software crunches data and helps guide sales employees. The software tries to predict which leads are likely to result in a deal, and it makes suggestions for the wording of sales pitches. The software gets into granular details like assessing whether it makes more sense to call or e-mail a particular target customer, and trying to predict the optimal time of day to contact people, Gorenberg says.
The product has shown the ability to increase customers’ revenue by as much as 30 percent, InsideSales claims.
“It’s allowing people who maybe weren’t as skilled in sales before, now to do better,” Gorenberg says. “I don’t think it’s going to destroy sales jobs; I think just the opposite.”
Gorenberg also sees opportunities for A.I. to lead to new jobs in construction, as cities remake themselves in an era of sensor-enabled infrastructure and self-driving cars. In manufacturing, data scientists and “sensor implementation workers” are being hired to help factory operations become more efficient and “smarter,” he says. And as education moves online and becomes more data-driven and personalized, the need for content creators and “teaching assistants” will increase, he says.
Gorenberg argues that in most cases, businesses are implementing A.I. technologies with the goal of improving their “top line”—meaning growing their revenue—rather than as a way to trim their bottom line by reducing human labor.
“I take the position that when a company is able to improve its top line, it gives it more money to spend, and [it] typically hires more people,” he says.
Lots of robotics and A.I. software companies these days are touting ways their technologies can aid human workers, rather than put them out of work. New York-based B12, for example, has promoted itself as “human-assisted A.I.” software. Boston-based Lola is developing a personal travel service that combines human travel agents with sophisticated software. On the hardware side, Boston-area companies like NextShift Robotics and Veo Robotics are trying to make it easier for robots to collaborate with humans.
“Robots and humans have really complementary strengths,” says Clara Vu, a co-founder of Veo and a former senior software engineer at iRobot. “Robots are strong, they’re fast, they’re precise. They can lift things that no human could ever lift. They can put them down with sub-millimeter precision. But dexterity, judgment, flexibility—humans just way outclass robots in that way.”
What that means, Vu continues, is it’s “not so much that a job can be automated. It’s pieces of a job can be automated. [As] machines get a little smarter, they’re not anywhere near approaching the intelligence of a human, but they’re getting intelligent enough that you can share some context.”
Vu sounds skeptical that machines will reach human levels of intelligence any time soon, if ever. She thinks it’s a problem that the general public’s understanding of robots is colored mostly by science fiction.
“I think that that’s very misleading,” she says. Rather, the advancement of robots and software is really just “another chapter in the story of us figuring out—us as humans, as people—figuring out new ways to build things and to make things. There’s change and adjustment.”
Gorenberg predicts that the term “artificial intelligence” will go away as machine learning algorithms and related technologies become ubiquitous in software systems. And when that happens, Gorenberg thinks the debate over “whether jobs will be created or lost” will also fade away.
He says it’s the same with most new technologies. At the beginning, “people fear it,” but in the end, it “produces positive results,” he says.
“I think we’re in a transition” period, Gorenberg says.