The 1970s Gave Us Industrial Decline. A.I. Could Bring Something Worse

 

 

“A silent recession has arrived for recent college graduates. Over the past two years, unemployment among 22‑ to 27‑year‑olds with bachelor’s degrees has climbed to levels seen during economic downturns. A college diploma used to be an ironclad job guarantee. Today it seems more like a lottery ticket with shrinking odds.

 

This plunge is just the beginning. As generative A.I. improves, entry-level and service sector jobs may increasingly disappear, threatening not just workers, but also the cities where they live. Recent research by the Brookings Institution shows how San Francisco and San Jose, Calif., New York and Washington could soon face significant job disruption, thanks to the rise of A.I. In San Jose, a striking 43 percent of workers could see A.I. transform half or more of their tasks.

 

There’s little evidence that A.I. has already begun taking jobs en masse. But just as manufacturing towns in the 1960s failed to recognize the looming threat of new technology, today’s leading service hubs risk underestimating the disruption of A.I. — especially as Silicon Valley races to automate white-collar work.

 

As the history of deindustrialization teaches us, spotting early warning signs is crucial to adaptation and survival.

 

Consider Pittsburgh in the 1960s, when it was still the steel capital of the world. Behind the scenes, a technological transformation was underway. Across the country, mini-mills began cropping up, employing far fewer workers than old blast furnaces did. Companies in Asia built enormous factories with refined technology, eroding U.S. Steel’s dominance. It took over a decade for the force of these advancements to be felt fully in America’s steel towns.

 

A similar story unfolded in Detroit. In the 1960s, new technologies — including robotics, computer-aided design and improved logistics — gave foreign manufacturers an edge and allowed car companies to operate smaller, more efficient plants. No longer bound to sprawling, unionized complexes along the Detroit River, the Big Three automakers expanded into the suburbs, then farther south and eventually to Mexico and Asia. By 2011, Detroit’s manufacturing employment had plummeted roughly 90 percent compared to what it was in 1950.

 

Modern service capitals tell themselves they aren’t dependent on one industry, yet diversification may be a thin cushion when the disrupter is a general‑purpose technology. Generative A.I. writes marketing materials, prepares tax returns, cleans data, codes software and drafts lesson plans.

 

In New York City the industries most exposed to those tasks — education, professional and technical services, information, finance and administration — together employ around 35 percent of the work force, a larger share than steel or autos ever commanded in their prime cities.

 

Skeptics note, correctly, that A.I. is still imperfect. But it does not have to replace workers entirely to have a big effect; it only has to drive down wages or make knowledge work easier and cheaper somewhere else. If a chatbot can help clients with paralegal work, market research or financial modeling, companies may no longer need to employ so many entry-level workers in Manhattan or San Francisco.

 

Like the manufacturing jobs of the 1970s, these roles could eventually move overseas. Coding assistants like GitHub Copilot boost developers’ productivity by 56 percent, with the greatest gains among beginner coders. Similarly, A.I. writing tools enable less-skilled writers, including nonnative English speakers, to create professional-quality documents. As location, linguistic ability and experience become less critical in determining productivity, the wages commanded by high-cost cities will encourage companies to hire in cheaper markets.

 

Deindustrialization’s lesson is not that decline is inevitable, but that reinvention is essential. Pittsburgh and Detroit didn’t collapse overnight. They stumbled because they failed to nurture new industries — in tech and professional services — as traditional jobs vanished. Throughout the 1970s and 1980s, Detroit and Michigan poured subsidies, tax abatements and infrastructure investments into the Big Three automakers, without ever halting their inevitable decline.

 

Boston, however, took a different path. It reinvented itself repeatedly: first in the early 1800s as a maritime center powering global trade; again in the late 1800s as a manufacturing hub fueled by immigrant labor; and in the late 1900s as a center of the tech economy and finance. Each leap depended on young talent and innovation, making education the engine of Boston’s repeated renewal.

 

So what kind of talent will matter in an A.I. age? No government can foresee exactly which sectors will spark the next wave of job creation, but cities can avoid propping up dying industries, as Detroit did, and instead seed new ones.

 

Innovation sprouts in environments that cultivate in-person interaction and experimentation — activities historically nurtured by urban areas, from Renaissance Florence to the cities that make up Silicon Valley. That means that governments should invest in amenities that attract and retain talented residents: public spaces, fast and affordable transit, top-tier schools, museums and theaters.

 

It also means making it easier for people to move from job to job and fostering competition. When Michigan began enforcing noncompete clauses in 1985, the flow of workers into emerging industries slowed as many left the state. California took the opposite approach; it restricted noncompetes, which allowed engineers at the once-dominant Shockley Semiconductor Laboratory to spin out Fairchild Semiconductor and later Intel, making Silicon Valley’s semiconductor miracle possible.

 

If local governments don’t act, they risk repeating Pittsburgh’s and Detroit’s decline. Knowledge-based roles pay significantly more than traditional assembly-line jobs, so they sustain a much broader network of local services, such as restaurants, retail stores and transit. The economist Enrico Moretti finds that while each manufacturing job supports about 1.6 local jobs, each high-skilled tech or professional role sustains around five. If even a small fraction of analysts, developers and paralegals move to cheaper regions — or are replaced by A.I. — it could have a profound effect on the whole community.

 

But if San Francisco, New York, Seattle and other service capitals lay the groundwork for innovation — embracing A.I. as a catalyst for new industries, rather than just as an efficiency tool — they could change their fate. Now is the time to start drafting the next chapter of urban prosperity.

 

Carl Benedikt Frey is an associate professor at the Oxford Internet Institute at the University of Oxford, and he directs the Oxford Martin School’s Future of Work program. His book, “How Progress Ends: Technology, Innovation and the Fate of Nations,” will be published in September.” [1]

 

 In the face of AI, cities really need to improve their creative potential. President Trump is overhauling US universities that have lost efficiency. New York is considering free public transportation. It's time for Vilnius to think about change too.

 

1.  The 1970s Gave Us Industrial Decline. A.I. Could Bring Something Worse: Guest Essay. Carl Benedikt Frey.  New York Times (Online) New York Times Company. Aug 19, 2025.