ROBOTS, AI AND LOST JOBS?
Automation will change the workforce — in some cases, it already has — but not as quickly as you might expect
One day, a software system could answer customer calls. In fact, there are already algorithms in chat bots that can translate languages in real time, and there are robots that take inventory in retail stores.
And that’s just what Pittsburgh companies are working on.
Technology is increasingly becoming an integral part of the workplace, acting more like a part of the team (except for maybe sharing lunch) and getting more tasks every day. As Pittsburgh establishes itself as a tech hub, it’s also finding itself at the forefront of an evolution likely to alter how just about everyone works — from the toll booth operator to the paralegal to the data analyst.
Automation could disrupt 25% of the workforce — about 36 million jobs — in the next few decades, based on predictions from a January 2019 Brookings Institute study.
By 2030, 20 million manufacturing jobs could be displaced by robots, according to a June study from Oxford Economics. Each new robot installed replaces an average of 1.6 workers, and 1.7 million jobs have already been lost since 2000, the study reported.
The goal is to use technology to make work better — to make jobs easier and safer, to make employees’ tasks more stimulating and creative, to make companies more efficient and productive.
But it could also leave some people behind.
“Companies recognize that at the end of the day, the people within the companies are core to what the company does,” said Anthony Gadient, co-founder and executive vice chairman of Voci Technologies, a company in the Strip District working on the software that could one day answer phone calls.
“A well-run company is going to make sure that they are investing in their team to ensure that as the technology evolves that that team is evolving as well,” he said. “Some companies, unfortunately, aren’t well run. In those situations, the change might be more traumatic.”
The software from Voci Technologies, which Mr. Gadient described as a “computer buddy,” listens along to a conversation and searches for ways to assist, like sifting through a database for the answer to a question or finding products to offer a customer.
Because the technology can understand the conversation, rather than just picking up on key words, it can learn as it listens.
So, in some instances, the software could actually answer the call, Mr. Gadient said. But it’s still a long way off from replacing the human in the call center.
“The reality is, the amount of intelligence that is needed today, you can’t go very far without needing to talk to a human,” he said.
That’s the way most technology is set up — it is designed to perform a specific task. Jobs that focus on things like baggage handling at airports, loading inventory in warehouses or preparing food at restaurants are likely to be among the first to go.
But technology won’t discriminate in its disruption of the workforce, based on findings from a November Brookings Institute report.
It has been well-documented that automation will disrupt blue-collar workers with manual labor jobs, such as truck drivers or pickers in an assembly line. The new study found artificial intelligence will also affect white-collar workers with routine tasks that could be replicated, like radiologists or computer specialists.
The augmented worker
HarbisonWalker International, a 150-year-old manufacturing company based in Moon, has opted to try both forms of tech disruption on its employees. So far, it’s worked.
The company, which produces refractory materials, has made a point to invest in new technologies. Martha Collins, the chief technology officer, said she even has one employee who goes out to “scout” what is being used in other industries to get ideas.
In October, the company unveiled a robot called the Mule-R that helps install its products in customers’ plants. It makes the installation process more efficient — especially important in a steel mill, which can go through a refractory product in a matter of hours.
HWI has also started using algorithms to maintain data about its plants and customers. Because the software does the heavy lifting, more employees are able to work with the numbers without needing a background in computer science or analytics, Ms. Collins said.
The manufacturer opened a “highly automated” plant in Ohio in spring 2018, and Ms. Collins said they are already seeing “more throughput per worker.”
Overall, employment has remained the same, she said. As machines, software and automated plants are worked into the process, people are moved around to different departments, a trend Ms. Collins predicts will play out in many different companies.
“People might not have to work as hard, which would be great,” she said. “I don’t think the number [of employees] is going to go down. I think the number is going to shift.”
Replacing Band-Aids with bots
That could be a welcome change for some manufacturing workers, said Stephen Catt, deputy director of education and workforce development at Advanced Robotics for Manufacturing, a national research institute based in Hazelwood.
Mr. Catt spent eight years working on a factory floor, a job that he said wasn’t sustainable. Most of his old colleagues now have health issues from walking miles on cement floors and rushing to meet deadlines.
“Everybody talks about safety protocols, but if you’re actually going to get your orders done on time, you can’t follow all of that. You just have to climb up like a monkey, put something on your shoulder and jump down,” he said.
At ARM, Mr. Catt said they are working to eliminate barriers preventing companies from keeping up.
A small business, for example, may not be able to spare a worker to complete a training course because they are responsible for 20% of the company’s output. Or, a worker who is 50 years old may not have the motivation to learn a new skill in the twilight of their career.
“Robots are taking dull, dirty and dangerous jobs away from our workers and usually those jobs are not jobs that anybody is going to miss,” Mr. Catt said. “But we need to transfer that passion and the willingness to work into these new technologies so that existing workers can adapt readily.”
On the factory floor, Doug Stanton, an engineer for ABB, is responsible for making sure people know how to use the automation machines he comes in to install.
He always gets a mixed response — some people are eager to learn what the new machine can do, while others are reluctant to let go of the old ones. Many of those are so outdated they are patched together with “Band-Aids,” Mr. Stanton said.
Mostly people are excited that it takes away some of the worst parts of their jobs.
“A lot of them look at it and say, ‘You mean I’m not going to have to wedge this up and make it run everyday now?’” he said.
‘For every job they take, they make new ones’
Bossa Nova Robotics, based in the Strip District, said its customers treat the bot just like an employee, going so far as to give it a nametag.
The autonomous robots travel around a store taking pictures of inventory and alerting employees when something is amiss or running low. It’s a problem the researchers who developed the tech didn’t know existed until they went to retailers pitching their new bot.
“They needed eyes on the shelf to know the real time state of the store,” said Sarjoun Skaff, co-founder and chief technology officer at Bossa Nova. “They see it as a productivity tool, much like the barcode scanning gun used to be in the past . ... That is a perfect utilization for a machine that has infinite patience.
“This way we allow people to do what they do best which is manipulate products and interface with customers.”
That is the argument many people make for why automation is valuable — it will free up time for humans to focus on more important and more exciting tasks.
It also opens up the possibility for the creation of new tasks no one has thought up yet.
Justin Starr, a professor of robotics at the Community College of Allegheny County, sees demand growing for people who know how to operate and maintain a robot but who don’t need to know how to build it.
“People think robots are coming to take jobs, and I think that’s true. I worry they’ll make a robot professor one day,” Mr. Starr said. “But for every job they take, they make new ones, because understanding how to work with these systems as these systems get everywhere is just a huge demand.”
Before going into teaching, Mr. Starr worked at RedZone Robotics, a company in Cranberry that created robots to inspect sewers.
He said it did cross his mind that these machines could one day take someone’s job, but the need to create something that could more quickly identify and remedy infrastructure problems outweighed any guilty feelings.
“It didn’t make sense to fight the robot,” he said of the people he trained. “They were coming whether they liked it or not, and there’s room for lots of people in that ecosystem.”
Slicing the ‘wealth pie’
At Carnegie Mellon University, researchers are working on illustrating a clearer picture of what the workforce will look like through its Future of Work initiative.
One project is testing how technology performs as a tutor in elementary schools. Another is examining the text of AI patents to determine the types of tasks developers are targeting.
Results are still preliminary, but Lee Branstetter, the head of the initiative, said they already show AI changes the types of workers companies are looking to hire.
“It’s increasing productivity, but, of course, that can be a double-edged sword — if firms can produce a lot more with fewer workers, then that can have ambiguous effects on labor and demand,” he said.
The fact that innovation grows the “wealth pie” through increased efficiency is important, said Tom Mitchell, a computer science professor at CMU who is also a part of the Future of Work Initiative. But, if left unchecked, that pie will grow more and more unequal on its own.
“AI generates a bigger wealth pie. Who could possibly be against that?” Mr. Mitchell said. “AI is also a force toward further skewing the distribution of that wealth. Who could possibly be in favor of that?”
“We’re really kind of flying blind into the future of jobs,” he said, adding that “we have a voice” in what the workforce will look like. “It’s not that technology is this steam roller that’s going to roll over us. We have some control over what we do.”
A robot made by Bossa Nova Robotics checks inventory in a mockup Wal-Mart aisle at the company's test lab on November 3, 2017 in the Strip District.
HarbisonWalker International's Mule-R, which helps install refractory products.