Geographic Analysis16 min readยท

The Rust Belt's Second Disruption: AI Hits Where It Already Hurts

Communities still recovering from manufacturing decline now face a new wave of automation. We analyze AI's impact on the Rust Belt's remaining employment base โ€” from call centers to healthcare administration.

The Rust Belt โ€” broadly encompassing Ohio, Michigan, Indiana, Pennsylvania, West Virginia, and portions of Illinois, Wisconsin, and New York โ€” lost over 4.7 million manufacturing jobs between 1979 and 2020. Communities that built their identities around steel, auto, and industrial production spent decades attempting to rebuild. Many pivoted to service-sector employment: call centers, healthcare administration, back-office processing, and distribution centers. Now, AI threatens these replacement industries with a cruel irony โ€” the jobs that replaced factory work are among the most vulnerable to the next wave of automation.

The Rust Belt's Current Employment Profile

Using BLS Quarterly Census of Employment and Wages (QCEW) data for the core Rust Belt states (OH, MI, IN, PA, WV), the region's employment landscape has fundamentally shifted:

SectorRust Belt Employment (2025)Share of TotalAI Displacement Risk
Healthcare and Social Assistance3,180,00018.2%Moderate (admin), Low (clinical)
Manufacturing (remaining)2,640,00015.1%Moderate-High (AI + robotics)
Retail Trade1,980,00011.3%High
Professional/Business Services1,750,00010.0%Moderate-High
Transportation/Warehousing1,120,0006.4%High
Government2,310,00013.2%Moderate (admin roles)
Financial Activities890,0005.1%High
Administrative/Support Services1,080,0006.2%Very High
Other2,530,00014.5%Varies

The Call Center Collapse

One of the Rust Belt's most important post-industrial replacement employers was the call center industry. Cities like Youngstown (OH), Flint (MI), Scranton (PA), and Huntington (WV) attracted call center operations with low commercial rents, available labor pools, and economic development incentives.

BLS data shows the region hosts approximately 320,000 customer service representative positions (SOC 43-4051), with wages averaging $35,200โ€“$39,800 in Rust Belt metros. These positions carry an ADI score of 74, making them among the most AI-vulnerable occupations.

The impact is already visible:

Metro AreaCall Center Employment (2022)Call Center Employment (2025)ChangeAI Attribution
Youngstown-Warren, OH8,4005,800โˆ’31%High โ€” major employer closures citing AI
Scranton-Wilkes-Barre, PA12,1008,900โˆ’26%High
Flint, MI4,2003,100โˆ’26%Moderate-High
Toledo, OH6,8005,200โˆ’24%Moderate
Erie, PA3,5002,600โˆ’26%Moderate-High
South Bend-Mishawaka, IN3,2002,500โˆ’22%Moderate

These communities, which spent years attracting call centers with tax breaks and workforce development funds, are watching those investments evaporate. And unlike manufacturing, which declined gradually over decades, AI-driven call center consolidation is happening in 2โ€“3 year cycles.

Healthcare Administration: The Hidden Vulnerability

Healthcare is the Rust Belt's largest employer, and it's commonly assumed to be AI-safe. Clinical roles largely are. But healthcare administration โ€” which accounts for roughly 30% of healthcare sector employment โ€” is highly vulnerable:

  • Medical coding and billing (SOC 29-2072): 44,800 positions in the region, ADI score 72. AI systems now code medical records with 94% accuracy versus 89% for human coders.
  • Medical transcription (SOC 31-9094): Already declining; AI speech-to-text has effectively automated this role. Regional employment down 55% since 2020.
  • Health information technicians (SOC 29-9021): 28,600 positions, ADI score 58. AI extracts, organizes, and validates health records.
  • Medical secretaries (SOC 43-6013): 62,000 positions, ADI score 65. AI scheduling, patient communication, and records management.
  • Claims processing (SOC 43-9041): 38,500 positions, ADI score 78. Insurance claims increasingly processed by AI end-to-end.

In total, approximately 210,000 healthcare administrative positions in the Rust Belt face high AI displacement risk โ€” a sector many displaced manufacturing and call center workers transitioned into.

The Warehouse/Distribution Dilemma

E-commerce growth brought distribution centers to the Rust Belt, leveraging central geography and interstate access. Amazon alone has opened 48 fulfillment and sortation centers across OH, MI, IN, and PA since 2015, employing approximately 95,000 workers.

But warehouse automation is advancing rapidly:

  • Amazon's Sparrow robotic picking system handles 65% of item types in equipped facilities
  • Robotic goods-to-person systems reduce picking labor by 50โ€“70%
  • AI-optimized routing reduces packing and shipping labor by 25โ€“35%
  • Amazon's per-facility headcount has declined an average of 15% since 2023 while throughput increased 20%

Regional warehouse and storage employment (NAICS 493) grew from 185,000 in 2019 to 245,000 in 2023, then plateaued at 248,000 in 2025 despite continued e-commerce growth. The growth phase is over; the contraction phase is beginning.

The Back-Office Brain Drain

Financial services, insurance companies, and corporate back-offices moved operations to Rust Belt cities for cost savings. These operations โ€” claims processing, mortgage servicing, accounts payable, data entry, basic financial analysis โ€” employ an estimated 380,000 workers in the region.

AI displacement in back-office functions is among the fastest-moving areas:

FunctionRegional EmploymentAI Automation LevelProjected 2030 Employment
Data Entry (43-9021)45,00085โ€“90%10,000โ€“18,000
Claims Processing (43-9041)52,00070โ€“80%18,000โ€“28,000
Loan Processing (43-4131)28,00065โ€“75%10,000โ€“16,000
Accounts Payable/Receivable68,00060โ€“70%28,000โ€“42,000
Basic Financial Analysis42,00050โ€“60%22,000โ€“30,000
Administrative Assistants (43-6014)145,00045โ€“55%80,000โ€“105,000

City-Level Analysis: The Most Vulnerable Communities

Metro AreaPopulationUnemployment (2025)High-Risk AI Jobs% of Total EmploymentResilience Score
Youngstown-Warren, OH538,0006.2%38,00016.8%Very Low
Flint, MI406,0007.1%26,00015.2%Very Low
Huntington-Ashland, WV-KY-OH356,0005.8%21,00014.9%Very Low
Erie, PA269,0005.5%16,00013.8%Low
Scranton-Wilkes-Barre, PA554,0005.3%35,00013.5%Low
Dayton, OH807,0004.8%48,00012.8%Low
South Bend, IN322,0004.6%18,00012.1%Low-Moderate
Toledo, OH641,0005.0%36,00011.8%Low-Moderate
Akron, OH703,0004.5%38,00011.2%Moderate
Detroit-Warren-Dearborn, MI4,340,0005.1%210,00010.8%Moderate

The resilience score accounts for educational attainment, industry diversification, proximity to growing metro areas, broadband access, and existing retraining infrastructure. Communities scoring "Very Low" have limited capacity to absorb a new wave of displacement.

Why the Rust Belt Can't Just "Retrain" Again

Standard policy responses focus on workforce retraining, but the Rust Belt's experience with manufacturing displacement reveals why this is insufficient:

  • The TAA Track Record: The Trade Adjustment Assistance program, designed for manufacturing displacement, has a completion rate of only 37% and only 58% of completers find employment at comparable wages (GAO, 2024)
  • Age demographics: The Rust Belt's workforce skews older โ€” median age 44.2 vs. national 42.0 โ€” reducing retraining ROI and employer willingness to invest
  • Educational infrastructure: Many Rust Belt communities have seen college enrollment decline; fewer training providers remain operational
  • Broadband gaps: 14% of Rust Belt rural residents lack adequate broadband โ€” a prerequisite for many remote/tech-adjacent jobs
  • "Train for what?": Unlike manufacturing-to-services, AI threatens the destination occupations of retraining, not just the origin occupations. What do you retrain a call center worker for when office work itself is being automated?

The Manufacturing-That-Remains Risk

The Rust Belt still employs 2.64 million in manufacturing โ€” down from peaks but still significant. AI compounds existing robotics-driven automation:

  • Quality inspection: AI vision systems replacing human inspectors โ€” 85,000 positions at risk
  • Production planning: AI scheduling and supply chain optimization โ€” 42,000 positions at risk
  • Maintenance prediction: AI predictive maintenance reducing technician headcount by 15โ€“25%
  • Administrative/clerical in manufacturing: 180,000 positions subject to same AI office automation trends

Potential Bright Spots

The picture is not entirely bleak. Several Rust Belt dynamics could be leveraged:

  • Energy transition: EV battery plants, renewable energy manufacturing, and grid modernization are creating new industrial jobs (Ford's BlueOval City in TN is a model; Ohio and Michigan are competing for similar investments)
  • Data centers: Cheap electricity, available land, and central location make parts of the Rust Belt attractive for AI data center development โ€” a category projected to add 100,000+ jobs nationally by 2030
  • Reshoring: Supply chain security concerns are bringing some manufacturing back from Asia; the Rust Belt's industrial infrastructure and workforce traditions position it for reshored production
  • Healthcare delivery: While admin jobs decline, clinical positions continue growing as the population ages. The Rust Belt's aging demographics drive healthcare demand.
  • Construction: Infrastructure spending (IIJA, IRA) disproportionately benefits the Rust Belt's crumbling roads, bridges, and utilities

Our Projection: The Rust Belt's AI Employment Impact

CategoryCurrent Employment2030 ProjectionNet Change
High-risk service jobs (call center, back-office, retail)1,680,0001,150,000โ€“1,350,000โˆ’330,000 to โˆ’530,000
Healthcare admin210,000140,000โ€“170,000โˆ’40,000 to โˆ’70,000
Warehouse/distribution248,000180,000โ€“210,000โˆ’38,000 to โˆ’68,000
Manufacturing (AI-specific displacement)2,640,0002,300,000โ€“2,450,000โˆ’190,000 to โˆ’340,000
Total high-risk displacement4,778,0003,770,000โ€“4,180,000โˆ’598,000 to โˆ’1,008,000

Against partially offsetting growth in healthcare delivery, energy transition, construction, and data centers (estimated +150,000 to +250,000 jobs), the net impact remains โˆ’350,000 to โˆ’760,000 jobs by 2030.

Conclusion

The Rust Belt is facing a second wave of automation-driven displacement, and this time the affected workers are the ones who already adapted once. Call centers, back-offices, healthcare administration, and warehouse jobs โ€” the very sectors that replaced manufacturing employment โ€” are now in AI's crosshairs. The communities with the fewest resources and the most fragile economies face the greatest exposure. Without targeted federal intervention that goes beyond recycled retraining programs, the Rust Belt's second disruption could be more damaging than the first โ€” because there may be no next sector to transition into.

Related Analysis