Worker Vs. Machine: Are Worker Fears or Expectations About Automation Realistic?
AI hasn’t come to many workplaces yet, but reactions are mixed where it’s being tested and implemented. Invested checks in with workers and experts in different industries to get a sense of how people are thinking about what AI means for their work.
Jay Lindsay is a writer and editor for Invested and Corporate Communications at the Federal Reserve Bank of Boston.
Just how big a threat are Flippy the hamburger-making robot and his kind to the American worker?
Flippy, a mechanical arm with a spatula hand, was hired last year with some fanfare by a California hamburger chain. A day later, Flippy was fired because he couldn’t keep up with burger demand, and he seemed like more show than substance. But fast forward to today, and an updated Flippy is flipping burgers in southern California.
Like Flippy, automated aides can appear gimmicky. But also like Flippy, the technology can ultimately prove effective, and it’s threatening when workers see machines performing jobs once reserved for humans.
Besides working at burger joints, automated machines—often assisted by artificial intelligence—are doing jobs as varied as driving downtown shuttles and writing high school football game stories. In 2018, union negotiators for striking Marriott hotel workers were so concerned about automated concierges, bartenders, servers, etc., that they insisted on—and won—protections from automation.
Courtney Leonard, a cocktail server at the Marriott-owned Westin Boston Waterfront Hotel, said coworkers initially dismissed the threat from robots, because robots can’t replicate the personal touch prized in the service industry. But she said it’s become clear the potential cost savings from automation are too alluring for companies to ignore. “My message (about automation) would be, don’t ever think, ‘It won’t happen here,’” Leonard said. “In one way or another, something is coming.”
At its core, automation creates ways for machines to complete tasks previously performed by humans. When aided by AI, it mimics human decision-making and adaptability.
Automation can be a real boon to humans when it takes over unpleasant or rote tasks, like data entry and tabulating, as it frees people up for more satisfying work. Flippy, in fact, got his first job because the hamburger chain was having trouble finding cooks who could tolerate constantly working over a hot grill.
But automation doesn’t just do things no one else wants to do. In Columbus, OH, the city proposed operating six new automated shuttle vehicles in an application for a federal program, and it explicitly cited their ability to replace human drivers. “A major benefit of a fully autonomous vehicle is the reduction in cost achieved by eliminating the operator and all onboard equipment necessary for human operation,” the application read.
The response by the bus drivers’ union was not subtle: “New Wave of AUTOMATION WILL DECIMATE GOOD Paying JOBS IN OHIO AND PUT TRANSIT RIDERS AT RISK!” read the headline of a pamphlet by Local 208 of the Transport Workers Union of America. The headline was placed over a picture of men standing in a Depression-era soup line.
Columbus’s first shuttle debuted downtown in December, when it ferried a maximum of six people at a time between four stops at a top speed of 25 mph. Columbus bus driver Darryl Neal said he knows more automated vehicles may be coming, but he doesn’t think it’s possible for them to adequately or safely replace human-driven vehicles.
Humans know which older customers need extra time or help getting on board, Neal said. They know from experience when a car might be looking to cut into traffic at the wrong time, so they can avoid the accident. They know that when a ball bounces in front of the bus, they need to brake immediately to avoid the child that often follows.
“We have a supercomputer in our heads,” Neal said. “We have a computer called the brain, and it has gut instinct, it has compassion and feelings for other human beings. … That’s why the operator needs to be behind the wheel of the vehicle, carrying other human beings.”
The news industry hasn’t seen the same level of pushback on automation. “Robot reporters” are writing news and sports stories for major outlets, including The Associated Press, which produces nearly 4,500 automated U.S. corporate-earnings stories a quarter, as well as automated NBA and NHL game previews. The Washington Post’s in-house automation program, Heliograf, writes stories on politics and sports, including stories about every high school football game in its region.
These robot reporters are really just sophisticated software programs that scrape data from sources like corporate earnings reports, then plug it into templates. In many cases, the story—say about a high school game in a remote area of Virginia—wouldn’t even exist if a robot didn’t take it on because it wouldn’t be worth the human investment, according to Jeremy Gilbert, who works in the Post’s newsroom as director of strategic initiatives.
Fears that automation could broadly replace human reporters are unwarranted, Gilbert said. Only humans can do in-depth interviews, stories that dig past the data, or colorful and nuanced writing. That’s what readers want and news organizations need during a time when the industry is relying more on subscription revenue and less on advertising revenue.
“I think it would be a mistake for newsrooms to say, ‘We are going to replace our human reporters with machine-generated content,’ and watch them lose out on the most engaged readers,” he said.
But even those who clearly see automation’s shortcomings know it can have a place and positive impact.
“Our approach was, we believe you aren’t going to stop technological innovation. That is an absurd premise,” said Carlos Aramayo, a vice president at UNITE HERE Local 26 in Boston who was involved in last year’s Marriott negotiations. “But obviously the way it’s used could be a positive or negative for the workforce. That’s the thing we really want to look at.”
Meredith Broussard, an AI researcher and data journalism professor at New York University, said people consistently overestimate the capabilities of technology and elevate it over human beings—she calls it “technochauvinism.” Underneath fears about an automated future is often “this kind of nutty vision of people being erased and replaced by machines, and it just being some kind of sci-fi situation,” Broussard said.
When it comes to how automation will affect the workplace, human beings have more power than they imagine, she said. “The algorithms aren’t suddenly going to become sentient and take over the world. We can all be certain about that. So then it comes down to human choices: What are humans deciding to do about other humans’ jobs?”
The views expressed are not necessarily those of the Federal Reserve Bank of Boston or the Federal Reserve System. Information about organizations, programs, and events is strictly informational and not an endorsement.
- Elements of AI - a free online course from University of Helsinki on the basics of AI
- Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy
- Computers and populism: artificial intelligence, jobs, and politics in the near term
- Don’t Fear the Robots: Why Automation Doesn’t Mean the End of Work
- Artificial Intelligence ,
- employment ,
- Automation ,
- Fear of technology ,
- Labor unions ,
Smart Places, Getting Smarter: Facts about the Young Professional Population in New England States
Labor Markets During and After the Pandemic
New England Study Group Past Meetings
Raising the Floor: Strategies that Make Jobs Work for All