A vacant home on February 7, 2012 in Youngstown, Ohio. —Benjamin Lowy—Getty Images

The story of Youngstown, Ohio, is often seen as a story about the past. But it’s time to start looking at the Rust Belt city as a warning about the future.

As the artificial intelligence revolution unfolds, debates about its impact on the American economy often hinge on the belief that if the technology creates more jobs than it destroys, it will be a success. But the more important question is often the most overlooked: Who will bear the costs of the AI revolution?

Youngstown is a symbol for what can happen when America stands idly by as market forces inflict abrupt economic transition on communities. And the next Youngstown may already be taking shape. It may not depend on a steel mill, but on concentrations of jobs increasingly vulnerable to AI. From Huntsville, Texas, to State College, Pennsylvania, the industries will be different, but the risk is familiar: jobs transformed or eliminated faster than the community around them can adapt.

For a decade, we have taught a Columbia Business School course called Bridging the American Divides. The class includes a four-day visit to Youngstown, Ohio, to help students understand what national economic statistics often fail to capture: what happens when a booming economy leaves a community behind.

Most Americans know the broad outline of Youngstown’s decades-long story. Once a proud and prosperous center of American steelmaking, its economic foundation collapsed as changes in steel production and foreign competition made much of the city’s industry uncompetitive. 

On “Black Monday” in 1977, Youngstown Sheet & Tube announced the closure of its Campbell Works, eliminating thousands of jobs overnight. Tens of thousands more disappeared in the years that followed. Youngstown’s population, which was over 120,000 at that time, has since fallen to fewer than 60,000 today.

While this shift hurt workers in Youngstown, many others benefited. From the perspective of the national economy, technological progress and global competition produced substantial benefits. As steel became less expensive, American automobile manufacturers and other importers of steel became more profitable and hired more workers. Consumers paid lower prices. And new industries and employment emerged in other American communities.

At the time, many economists highlighted those gains for the country. They weren’t wrong. But they largely missed something equally important: the costs were not spread across America. They were concentrated among specific workers, families, and communities. 

Over the past decade, our students have met former steelworkers, entrepreneurs, educators, nonprofit leaders, clergy, journalists, union heads, and local officials in Youngstown. They have seen that when a dominant industry in a community disappears, what is lost extends far beyond paychecks.

Jobs disappear, but so do tax revenues, local businesses, and confidence in the future. Neighborhoods can become ravaged by crime and drugs. Young people typically leave. Those who remain tend to not only lose employment but also identity and purpose. Distrust grows toward political, economic, and educational institutions that seem unable or unwilling to help. The result: community and social bonds are diminished.

Artificial intelligence promises another economic leap in increased productivity and lower costs. As AI reshapes employment in customer service, accounting, software, legal support, insurance processing, and other fields, the consequences will be felt disproportionately in communities whose employment is concentrated in those occupations. 

A Brookings report finds “the concentration of exposed and vulnerable workers is greatest in smaller metro areas” like Laramie, Wyoming, Huntsville, Texas, and Stillwater, Oklahoma. The Federal Reserve Bank of Philadelphia identifies New Jersey’s Trenton-Princeton metropolitan area and State College, Pennsylvania, as highly exposed. MIT’s AI Labor Map identifies hundreds of cities with “economic exposure” from Tallahassee, Florida, to Boulder, Colorado.

We can also expect new industries to emerge. Different jobs will be created in other places. Our prediction: the average national employment statistics will be just fine.

But history shows that new opportunities do not necessarily reach the workers and communities that lose jobs. If jobs transfer to new places, some people may not be able to move with them—they have mortgages, aging parents, children in school, social networks, and identities rooted in place. And many workers have developed skills over decades that are not easily transferable. Labor markets adjust faster than communities do.

It’s time for economists, business leaders, and elected officials to start asking whether the workers and communities disrupted by AI will have realistic pathways into gainful employment and the support needed to weather the transition.

Youngstown, in its efforts to progress, can provide guidance for communities facing AI disruption.

The city today is not simply a symbol of decline. We have watched entrepreneurs at Youngstown’s Startup Business Initiative launch new businesses, Youngstown State University build workforce pipelines for emerging industries, nonprofits like Youngstown Neighborhood Development Corporation rebuild neighborhoods, and Flying High rebuild lives, and civic leaders attract new investment for GM’s transformation of their massive Lordstown factory. Progress has been real, if uneven. The first lesson of Youngstown is not that communities cannot recover. It is that recovery rarely happens through market forces alone.
For communities to recover, institutions need to work together. After several decades, Youngstown is moving forward with sustained collaboration among business, government, educational institutions, nonprofits, and local leaders. Each plays a different role. Businesses invest and create opportunity. Governments build infrastructure and fund development. Educational institutions prepare future workers for new industries, and community organizations rebuild blighted neighborhoods and provide critical support. The second lesson is that recovery is a collective effort, not the work of any one institution.

Finally, recovery takes time, so there can be no expectation of magic short-term revitalization. That may be the most important lesson Youngstown offers as artificial intelligence reshapes the American economy. The goal should not be to slow technological progress. America has always prospered by embracing innovation. The goal should be to ensure that the gains from AI will go towards supporting the businesses, governments, educational and nonprofit organizations that must work together to rebuild the communities that bear its costs.

The next Youngstown does not have to happen. Avoiding it will require asking a better question now, before AI transforms the economy; not simply whether America will benefit, but whether the communities disrupted by that progress will have the opportunity to share in it.