It’s a thrilling time for technology, with innovations in artificial intelligence and robotics propelling society ever faster forward. But is it too fast?
That’s a question that came up more than once Wednesday night during a panel discussion on automation’s impacts on society and work, presented at Seattle University by the MIT Enterprise Forum of the Northwest.
One questioner noted that past industrial revolutions have taken place on the scale of a century to a half-century, but that the current workplace revolution is unfolding on a much shorter timescale.
“Institutions cannot deal with this rate of change, nor can most people,” he said. “How do you propose … at least trying to make the rate of innovation slow enough that everybody can catch up?”
Cesar Keller, a former Microsoft marketing executive who now heads a Bellevue, Wash.-based management consulting firm called Collective Brains, turned the question on its head.
“The only constant in the world is change, right?” he said. “Most of our population actually got to a status where you say, ‘OK, I don’t need to change all the time. That’s great: I have a profession, I have a paycheck, maybe I can retire in that profession.’ And things are stable.
“That reality is gone,” Keller said. “There’s no way back.”
The questioner persisted. “Yeah, but a 55-year-old, with 25 or 30 years in an industry, just can’t say, ‘Oh, let me go learn a new trade,'” he said.
Isn’t there a way to moderate the rapid rise of automation?
“I don’t believe so,” Keller answered. “We need to help them transition, or live inside that. … There’s no way back.”
Over the long run, economists say AI and other technologies could well create more jobs than they eliminate, as has been the case with past industrial revolutions.
But in the short run, there could be a painful transition ahead. “What is not fun is leading that transition,” Keller said.
One study suggests that half of America’s jobs are vulnerable to automation. Another study says 38.6 million Americans may have to switch occupations or learn new skills to hold down a job in 2030.
Jobs that involve repetitive tasks, such as serving up meals at a fast-food restaurant, will be especially ripe for replacement, said Ece Kamar, senior researcher in the Adaptive Systems and Interaction Group at Microsoft Research.
Also, if there’s lots of training data available for a particular skill — whether it’s playing the game of Go or driving a truck — that opens bigger opportunities for AI.
The path to automation’s rise also could bring tragedies worse than the loss of a job, as demonstrated by a fatal traffic accident in Arizona in which an Uber autonomous vehicle struck a pedestrian.
The incident prompted Uber to suspend its self-driving vehicle tests nationwide.
“Our focus so far has been cooperating with investigators and reaching out to the victim’s family and seeing what we can do,” said Caleb Weaver, who manages public affairs for Uber in Washington state. “I do believe the results of the investigation will really determine what the long-term impact of this is.”
For now, Uber’s view is that autonomous vehicles will transform transportation, but not necessarily lead to massive job losses — at least when it comes to personal transportation and short-haul trucking for deliveries. “There will be probably a decrease in the need for long-haul truckers, because that is the first, most obvious place for automation,” Weaver acknowledged.
“At least on the transportation side, it is going to be a far more complex and long-term transition that happens,” he said.
Weaver and others on the panel emphasized that robots and AI programs are more likely to augment what humans do in the workplace, rather than taking them totally out of the loop.
“Our belief is that there’s this notion of augmentation, where you’ve got a human and a machine working together, and together those create sort of a superhuman,” said Aaron Reich, senior director of innovation at Avanade, a global professional services company headquartered in Seattle.
In that vein, Kamar pointed out that AI-based, second-opinion software has been shown to reduce the human error rate for breast cancer diagnoses by 85 percent.
Dan Liebling, staff software engineer at Google Research, said the human-plus-machine formula is in line with a Hippocratic Oath for AI practitioners, which was inspired by a Microsoft book titled “The Future Computed” and drafted by Oren Etzioni, CEO of Seattle’s Allen Institute for Artificial Intelligence.
“My AI will seek to collaborate with people for the greater good, rather than usurp the human role and supplant them,” the oath reads.
Liebling seconded that emotion.
“Put humans before machines,” he said. “It’s not about technology, it’s always about humans.”
