The Robots We Were Afraid of Are Already Here

Digit drew a crowd, even here, in a convention center full of robot aficionados.

A humanoid warehouse worker, Digit walked upright on goatlike legs and grabbed bins off a shelf with muscular arms made from aerospace-grade aluminum. It then placed the boxes on an assembly line and walked back to the shelf to search for more. The crowd, which had assembled at ProMat, the premier trade show for the manufacturing and supply chain industry, held up phones and watched, a little quiet, wondering if at some point the robot would teeter and fall. It did not.

Digit, made by Oregon-based Agility Robotics, is the kind of technology that people have worried about for generations: a machine with the strength and adroitness to rival our own, and the ability to take our jobs, or much worse. Then ChatGPT came online, and suddenly the fear was of something smarter rather than stronger — malevolent bots rather than metallic brutes.

The automaton is still coming. It might not be ready to take over the Amazon warehouse yet, but the long-anticipated robot revolution has begun, accelerated in large part by the pandemic and the thunderous growth of e-commerce. Machines like Digit are ready to take over a vast swath of physical labor, from operating forklifts to doing the laundry.

Ron Kyslinger thinks this is a good thing. Mr. Kyslinger, an engineer who has spearheaded automation for some of the largest retailers in the world, including Amazon and Walmart, is passionate about the potential of robots to improve the quality of life for workers. Robots free humans from boredom, repetition, physical strain and productivity limits that can put their jobs at risk, he believes. He also believes that Americans have a prejudice against automation because of movies like “The Terminator,” inhibiting them from adapting to technology in ways both beneficial and inevitable.

Mr. Kyslinger, 56, is currently a consultant for companies hoping to increase automation, and his services are in high demand. Known for his ability to see the big picture not just in a warehouse full of whirring machines but across the global landscape of automation, he is blunt and methodical, and can be somewhat robotic himself in his personal manner. He is often hired to diagnose problems and tell a board or chief executive how it really is.

And how it really is right now, in Mr. Kyslinger’s opinion, is that the world is on the brink of enormous changes when it comes to the presence of robots at work.

“I don’t think people really understand where we are,” he told me. “We’re just scratching the surface.”

Use of robots by big brands, retailers and movers of goods accelerated significantly after 2019. According to the Association for Advancing Automation, robot orders in North America jumped 42 percent during the pandemic after essentially being flat over the previous five years.

The shift has taken place largely out of sight, inside an archipelago of windowless warehouses across the Southeast and Midwest, helping companies to avoid inflaming the taboo against replacing human workers with machines. Some are reluctant to even discuss automation.

Americans have long felt ambivalent toward automation. The country that invented such job-killers as the dishwasher and the combine also produced the likes of Philip K. Dick and James Cameron, artists whose dystopian visions helped breed lasting anxiety toward robots.

Over the last few years, significant resources have been thrown at making robots profitable — and this is paying off. More companies are competing to solve the problems that have traditionally come with automation, and many are succeeding.

“People are finally making money,” said Samuel Reeves, chief executive of FORT Robotics, a Philadelphia start-up focused on robot safety. “You’ve got legit work being done by mobile autonomous robots. And that’s only in the past two or three years.”

Interest skyrocketed during the pandemic and helped create relatively low-cost automation systems that companies can install very quickly, said Ash Sharma, managing director of Interact Analysis, which surveys trends in automation.

“We’ve seen billions of dollars flying into this sector,” he said.

Yet the use of robots in most countries remains relatively low, suggesting that a true reckoning with their social impact lies ahead.

The United States ranked ninth in robot density in 2021, down from seventh in 2020, according to the International Federation of Robotics. By contrast, in East Asia, where the aging of populations has long bred fears of worker shortages, robots have been embraced.

The United States has its own problems with aging workers, however, especially in heavy industries like manufacturing, where baby boomers form an outsize part of the work force.

“The pandemic took somewhere between one and a half to three million people out of work,” said Joseph Campbell, senior marketing manager for Universal Robots. “A lot of boomers who were planning to work past 65 said 62 is good enough. It’s scary.”

If a transition to a robotic work force is underway, managing it is likely to fall to a small group of industry veterans such as Mr. Kyslinger. Almost 20 years ago, he was one of a small number of robot boosters who built the paradigm, and to see it today through his eyes is to see where it may be headed next.

He has his concerns — about people, not robots.

In it for ‘what’s best for humanity’

At ProMat, which took place over four days in March inside Chicago’s McCormick Place convention center, Digit was the undisputed star. The scene looked, for the most part, like a kid-friendly science fiction movie, a robot bazaar where the machines move slowly, say excuse me and execute limited tasks like picking up objects and dropping them.

However, some robots sat inside plexiglass cages. “You don’t want to go in there,” Mr. Kyslinger said, pointing at one. “That thing will knock you on your butt.”

ProMat’s 51,000 attendees — a glad-handing throng of well-groomed, middle-aged white male faces attached to monogrammed backpacks and fancy sneakers — ambled from one exhibit to the next like visitors at a zoo. The crowd included buyers from major retailers and consumer goods companies, as well as venture capitalists and engineers.

At one booth for a robot “picker,” I told one of the few women seemingly within miles that I was writing an article about a guy named Ron. “Oh, really,” she said, looking fatigued. “There are a lot of Rons here.”

Mr. Kyslinger floated through this milieu like a celebrity entering a restaurant, barely able to move without being accosted. “Welcome to Ron’s world,” one attendee whispered to me above the whine of micro-motors. “Ron’s the O.G. of automation,” another said. Mr. Kyslinger blushed at the attention.

“I don’t love talking about me,” he said a bit gruffly. He had agreed to share his passion for and concerns about automation in the interest of “what’s best for humanity.”

Mr. Kyslinger, who grew up in western Pennsylvania and was a curveball-throwing right-hander for the University of Pittsburgh, majored in computer science in college. Practice was at 5 a.m., so he got up at 3:30 and went to the computer lab.

After graduating in 1989, he went into the automotive world, where, starting as a controls engineer, he spent 23 years working for Chrysler, Ford and Honda. Car companies were among the first to embrace automation, replacing humans with crude, often dangerous robots on assembly lines.

In the 1980s, robots were a rare example of the U.S. car industry’s using technology to think ahead. But in some ways, industry experts believe, they set robotics back.

“Everything we sold was to take labor out,” said Mr. Campbell of Universal Robots. “Everything was to replace a worker. That was the impression, and at that point it was the truth.”

For Mr. Kyslinger, who today lives near Myrtle Beach, S.C., working for a Japanese car company was a formative experience. He admired what he saw as Japanese culture’s disciplined approach to complex problems and wrote a master’s thesis on the different working environments at Honda and Ford.

In 2011, Mr. Kyslinger moved to an industry trying even more aggressively to automate industrial workspaces: food distribution. At C&S Wholesale Grocers, the nation’s largest grocery distributor, he designed a warehouse in which robots traveling 30 miles per hour filled up pallets destined for supermarkets.

C&S is a little-known company that has a hand in transporting an outsized portion of the nation’s food. (“If it belongs on a supermarket shelf, it’s probably moving through a C&S warehouse right now,” the New Hampshire-based company likes to say.) Under Mr. Kyslinger, C&S pioneered warehouses with so few human workers that they came close to the industry goal of “lights out,” meaning the ability to operate in darkness, minus human eyes.

Today Mr. Kyslinger says lights out is “getting really close.” As we walked the floor of Promat, he identified robots that were closing the gap with humans and in some cases outperforming them.

On the whole, Mr. Kyslinger is not easily impressed. He called the robot trade show full of “bits and pieces of the future — many parts make my brain hurt.”

Nevertheless, somewhat begrudgingly, he pointed out some highlights: A robotic arm with a kind of gripper that approached the versatility of human fingers. A visual sensor that had made progress in discerning the glare on a plastic wrapper from an object it contained. A sorter that excelled at finding the ideal geometry inside a cardboard box for items of different shapes, whether “toothpaste, tuna fish or a teddy bear.”

Mr. Kyslinger seemed more interested in explaining what robots can’t do. As a former computer programmer, he was decidedly unfazed by ChatGPT, the chatbot powered by artificial intelligence.

“People are looking for it to be the Holy Grail,” he said, “but it’s only as good as the people who programmed it.”

At another booth, we watched a “cobot” — short for collaborative robot — as it placed items inside a droid-like “autonomous mobile robot” zipping to and fro. Cobots are supposed to be harmless, but this machine’s Japanese manufacturer, Fanuc, a longtime provider of robotic arms to carmakers, had put it inside a cage anyway.

Its writhing octopus-like appendages were hypnotic, but as soon as we started watching it, one of its “end effectors” (robot-speak for hands) failed to grip a box of Q-Tips and dropped it on the floor, where the autonomous mobile robot ran over it.

“See?” Mr. Kyslinger said. The cobot was shut down so an attendant could enter the cage and remove the now-flattened Q-Tips. Mr. Kyslinger called this not a defect but an “edge case,” the kind of routine mistake that makes human intervention inevitable. It showed why getting to “lights out” in e-commerce will be a struggle.

“They probably taught the robot how to do this thousands of times in their lab,” he said. “Then they bring it out here and it still misses.”

We walked on, with Mr. Kyslinger being greeted at every turn. At one booth he stopped in his tracks. By now we had passed robotic lift trucks stacking boxes, robot sentry dogs with thermal and acoustic sensors, and robots unloading trailers. (At Automate, another industry event, we watched a Fanuc robotic arm twirl around 2,200-pound objects as if they were pizzas.) None of those moments, however, really made Mr. Kyslinger’s pupils widen.

“That catches my eye,” he suddenly said. It was a Shelby Cobra, made in 1967 and retrofitted for racing. A car.

‘Fear of our own creation’

One of Mr. Kyslinger’s many consulting clients wandering the ProMat floor was Samuel Reeves, a roboticist from Philadelphia. Mr. Reeves, now 40, began working on a company he called Humanistic Robotics in the mid-2000s, shortly after he graduated from college. It was devoted to land mine removal, the kind of extreme task that robots have long been assigned. Humanistic Robots used a construction vehicle to build a 10,000-pound minesweeping robot that could move on its own.

“And we were immediately terrified of it,” Mr. Reeves said. He then founded FORT Robotics, “born of our fear of our own creation.” He expressed the same kind of terror and regret that artificial intelligence creators have been voicing.

FORT Robotics is a “robotic-controlled platform,” according to Mr. Reeves, and one of only a few companies focused on preventing robots from mauling workers, which Mr. Reeves called “a disaster waiting to happen.”

“In the last generation of automation, people were just really trying to get machines out there that worked,” Mr. Reeves said. “There have to be huge innovations to improve safety so that machines can run autonomously and faster around humans — and at a lower price point.” He added that “safety-rated scanners are incredibly expensive — like $10,000 a pop.”

Mr. Kyslinger echoed Mr. Reeves’s concerns.

“I’ve seen robots go horribly wrong,” he told me. In one warehouse owned by a company he advised, a robot essentially clobbered a worker, breaking several bones. A technician had accidentally disabled its safety features.

“Human error causes problems, not robot error,” Mr. Kyslinger said, noting that airplane crashes have declined sharply since autopilot was introduced. “The robot does what it’s told to do — no more, no less.”

“People think of ‘The Terminator,’” he added, “but that stuff can’t happen when you have safety protocols.” Such protocols can include “bifurcating” a robot’s safety controls so that two humans would have to agree to the sort of change that would put workers near the robot at risk.

Safety concerns have made cobots one of the fastest-growing segments of industrial automation. A cobot “can hit you, but it can’t hurt you,” Mr. Kyslinger said. “It knows you’re there. It senses you’re there and stops.”

Robots show up

In 2018, before the pandemic unleashed a torrent of pressure on companies to automate, the Massachusetts Institute of Technology commissioned a task force on “The Work of the Future.”

The task force concluded that “no compelling evidence suggests that technological advances are driving us toward a jobless future.” On the contrary, it anticipated “more job openings than workers to fill them.” Nevertheless, it argued, “the implications of robotics and automation for workers will not be benign.”

“The pandemic laid bare vulnerabilities that have come from hollowing out U.S. manufacturing capabilities,” said one of the task force’s co-chairs, Elisabeth Reynolds, an M.I.T. lecturer who went on to serve as a special assistant for manufacturing and economic development on the National Economic Council. “Automation is going to help us make the transition to an advanced manufacturing center, while helping with a long-term shortage of workers.”

Another co-chair, Prof. David Mindell, agreed, calling the recent changes potentially positive for workers but only if the widespread incorporation of robots leads to the creation of “new industries and new kinds of jobs.”

“Sixty percent of the jobs in the Department of Labor database did not exist in 1940,” Mr. Mindell said over Zoom. “You know, web designer, massage therapist, dog walker, aerodynamic simulations engineer. We have to make sure that we’re continuing to create those kinds of jobs.”

At ProMat, Mr. Kyslinger and I eventually came to Digit, the humanoid warehouse worker. We watched Digit work itself into a crouch so it could pick up a bin near the floor.

“You don’t want people bending over to lift from down there,” Mr. Kyslinger said. “That’s where injuries occur, in their backs, in their necks.”

Still, the machine was moving slowly — slower than most humans. Mr. Kyslinger studied its movements, seemingly unimpressed. “A lot of algorithms go into that,” he said. “Humans do those things without even thinking.”

But, he added, to be an improvement over humans, machines don’t need to be faster.

“Robots show up every day,” he said.