Technological disruption often has a distinct geography: industrial automation and the Midwest, for example. And the landscape of generative AI’s impact is starting to come into focus, with two interesting analyses out this month.
It will have a big skyline.
The first study, from the Brookings Institution, used OpenAI data to identify the cities and counties most likely to see their jobs impacted by AI—for good and for ill. Not surprisingly, places like San Jose and Seattle, which are incubators for the technology, have a high likelihood of being reshaped by generative AI. But so do places like Durham and Dallas.
“The greater the education and pay level, including the more office- and information-oriented the place is, the more exposed it is,” says Mark Muro, senior fellow at policy director of the Metropolitan Policy Program at Brookings.
The second analysis, put out by OpenAI itself, looks at the adoption of ChatGPT among college-aged adults in states across the country. California, Arizona, New York, New Jersey, and Virginia—all with big tech industries—are the top five for adoption, while America’s Heartland, save Illinois, lags behind.
The Big Idea: The studies start to paint a picture of the geography of generative AI—one in which workers in broad swaths of the middle of the country may be spared disruption, but also are at risk of being left behind.
“The full impact is still unfolding,” the OpenAI report says. “But state-by-state differences in student AI adoption could create gaps in workforce productivity and economic development, impacting US competitiveness.”
It’s a concern Muro shares. “This technology could be extremely helpful for rural and smaller town communities, but it can’t be if there’s not much involvement with it,” he says.
Winners and Losers: Of course, being less exposed to AI also means those communities may be spared the most wrenching disruption from the technology.
A growing body of research indicates that generative AI could widen inequality in the nation’s cities, though that is still very much an open debate. A recent study from Brookings, led by Molly Kinder, found that the people most likely to have their jobs automated away—not just augmented or altered—are the 19M professionals in office support and administrative jobs that constitute a big chunk of the first rung of the middle class. They are predominantly women.
Meanwhile, the work of architects, engineers, and high-tech professionals—who are more highly paid—is likely to be remade, but those jobs aren’t expected to go away. Neither will many well-paid healthcare and management jobs, though all so-called knowledge work may face downward wage pressure.
The AI transformation for the nation’s cities could be bumpy, Muro says, and if not handled properly could leave some workers worse off.
“There’s no free productivity here,” he says. “There’s no way this won’t be disruptive.”
What to Do: The response will have to be nuanced, Muro says. Larger cities need to focus on this as soon as possible. A two-tracked response should make sure they are capitalizing on the upside—including making sure the benefits are widely shared—and also preparing for disruption.
And states need to be thinking about how to make their urban centers competitive nationally, while also ensuring that smaller cities and rural areas have exposure to AI.
Muro says new kinds of education programs, including reskilling and upskilling for incumbent workers, will have to be part of the equation. But it’s hard to know at this point what skills will be durable or newly in-demand.
“We keep saying we need to work much harder on workforce training and support,” he says. “But providers are really struggling to understand what those so-called durable or human skills are.”
On the Ground: The Milwaukee metro area, with about 1.6M residents, is among the areas of Wisconsin that are most exposed to the impacts of generative AI. Nationally, though, it ranks about average in the Brookings research, and it’s in a state with a relatively low expected impact. Wisconsin and the region, though, are leaning into generative AI and other forms of artificial intelligence.
That’s due in part to the state’s demographic reality, with an aging and shrinking population, says Kathy Henrich, CEO of MKE Tech Hub Coalition, who sat on the state’s legislative study commission on AI. “Without some of this AI automation, we’re actually going to have a gap in our ability to grow the economy,” she says.
Milwaukee is investing in AI learning labs for K-12 students. And local colleges—including the Milwaukee School of Engineering and Waukesha County Technical College—were early to develop degrees, certificates, and concentrations focused on AI. And the Northwestern Mutual Data Science Institute is a major hub of research and workforce development.
The MKE Tech Hub is creating AI roadmaps for startups and manufacturers—with a particular focus on being the “entry point and navigator” to help midsize manufacturers incorporate various forms of AI tech and train up their workers. The approach to education and training is proactive—rooted in a belief that preparing workers on AI can help drive adoption and economic growth.
“People are going back and getting these skills to go back into the businesses before they’ve been let go,” says Henrich.
The Kicker: Without smart policy and investment, most mid-size manufacturers wouldn’t know where to begin. “We tend to think that the market will solve itself,” Henrich says. “The market will not solve itself here. It’s going to need some assistance.”