The AI boom won't follow the same path as the US's manufacturing decline, an economist says. Workers will have a better chance of finding new jobs.
AP Photo/Rick Bowmer
One of the big questions facing US economists is what the coming AI revolution will mean for workers.
In the years ahead, there are concerns that companies' huge AI investments — and their ongoing adoption of these technologies — could eliminate jobs. A Goldman Sachs report published last year estimated that 300 million full-time jobs across the globe could be disrupted — not necessarily replaced — by AI in the coming years. It's already costing some workers their jobs.
Similar to US manufacturing workers who lost their jobs in recent decades to advancements like automation, those displaced by AI could find themselves without the skills needed for the modern workforce.
But there might be some reason to avoid a doom and gloom outlook. According to Steven Davis, senior fellow and director of research at the Hoover Institute — a public policy think tank based at Stanford University — the average person who loses their job due to AI will be better positioned to find a new role than the average US manufacturing worker who lost their job in recent history.
"For any given scale of job loss caused by AI, there's going to be less economic hardship than we can anticipate from our previous experience with industrial job loss," Davis told Business Insider.
In 2000, roughly 17 million Americans worked in manufacturing, but by 2010, the number had fallen to about 11 million — it's roughly 13 million as of June. Much of the decline was in the Midwest. Between 1990 and 2019, employment in the Midwest's manufacturing industry fell by over 1 million, or about 21%.
Various factors, including automation and globalization, have driven the decline, and many workers who lost their jobs struggled to find work in the aftermath. However, Davis said workers who are displaced by AI should have a better chance of avoiding the same fate. It comes down to when — and where — AI job losses are likely to materialize.
To be sure, there is significant disagreement among experts about what the AI future will mean for US workers. But some say looking at previous disruptions to the labor market could provide some insight into what lies ahead.
Some level of job loss is inevitable in the US economy — whether it be due to a recession, automation, or some other factor. In June, over 1.4 million Americans were laid off or discharged, among the lowest monthly figures over the past two decades. But when a person loses their job and can't find a new one, a temporary setback can end up having a lasting impact on their finances.
Therefore, when one weighs the impacts of the coming AI boom, Davis said it's wise to consider not only how many jobs it could displace — which happens to be extremely difficult to forecast — and focus on a more manageable question: How hard of a time will any displaced workers have finding new roles?
Davis said there are several reasons workers who are displaced by AI should have an easier time finding work than many manufacturing workers of the past.
First, when manufacturing jobs have been lost, it's often happened in bunches. A factory closed or a recession led to widespread cuts — Davis said economic downturns have historically hit the manufacturing industry "especially hard."
Not only have many manufacturing workers experienced job losses around the same time, but they've also tended to be clustered geographically around large manufacturing facilities, Davis said. This created situations where a large number of laid-off manufacturing workers with similar skills were looking for jobs at the same time in the same area — and there often just wasn't enough work to go around.
"If you enter the market or you need to find a new job in a recession, it's typically harder, not just because there's less hiring going on but because there are lots of other people looking for jobs at the same time," Davis said.
However, job losses in most industries, including those expected to be impacted by AI, are less likely to happen at the same time and in the same area, Davis said. So even if AI displaces some jobs, he expects workers to be in a better position to find a new role.
Additionally, to the extent workers displaced by AI are concentrated geographically, Davis said they're more likely to be in densely populated areas like cities. Impacted workers in cities would be more likely to have job opportunities than workers where manufacturing jobs were concentrated, which often were in the Midwest.
Secondly, the rise of remote work has also led many companies' workforces to be more geographically dispersed than they used to be. So even if a company replaced many of its workers with AI, these people would be less likely to compete with each other as they looked for new jobs.
"For the people who lose their jobs, at least many of them are not in a local labor market that's being hit by a wave of job losses from a single company or a single industry," Davis said. He added that remote working arrangements could also expand job opportunities for impacted workers — they wouldn't only have to apply for jobs in their local areas.
In June, about 24% of full-time US workers age 16 and older worked from home at least some of the time, according to the Bureau of Labor Statistics — 11% were remote for all their working hours.
While Davis is uncertain about the timing and scale of AI job displacement, he said he doesn't expect AI to drive major job losses over the next decade.
"I think AI is a potentially revolutionary technology, but it's one thing to develop the technology itself," Davis said. "To actually commercialize it at scale takes a long time typically, and we'll run into political resistance and policy resistance that will also slow down the process."
Experts have raised concerns about AI's potential to produce misinformation, violate copyright law, and even lead to human extinction. Congress has yet to pass comprehensive regulation related to AI, but President Joe Biden signed an executive order last year that included safety standards for AI researchers to follow.
Davis said that historically, it's taken decades for new technologies to be adopted across the entire economy. He pointed to electrification, the microprocessor, and the steam engine as past examples.
While there's been speculation that underlying AI technologies could improve faster than previous technologies, Davis said he thinks adopting them at scale will remain a challenge.
He isn't alone in his skepticism. Jim Covello, Goldman Sachs' head of global equity research, expressed a similar sentiment during a "Goldman Sachs Exchanges" podcast released on July 16.
"We're a couple of years into this, and there's not a single thing that this is being used for that's cost-effective at this point," he said of AI technology. "I think there's an unbelievable misunderstanding of what the technology can do today. The problems that it can solve aren't big problems."
What's more, Davis said there's also plenty of potential for AI to create jobs, adding that the early evidence suggests this tech isn't displacing jobs — but simply changing them.
"It's different than introducing a bunch of robots into a manufacturing plant and saying we don't need the workers anymore," he said. "That may eventually happen with AI, but that's not the main way it's been playing out so far."
However, in the decades ahead, Davis said it's possible that AI could lead to wider job displacement. He mentioned customer service agents as one profession that could be particularly vulnerable.
"The big question is does it happen gradually over time, in which case it would be relatively easy to accommodate," he said.
Have you lost your job due to AI? Have you used AI tools to save time at work? Share your experience with this reporter at jzinkula@businessinsider.com.
