While many experts warn that artificial intelligence could devastate white-collar workers the same way Chinese import competition hurt factory workers in the 2000s, a new analysis from Economic Innovation Group researchers Adam Ozimek, Jason Harrison, and Nathan Goldschlag published June 18, 2026 concludes the opposite. The "AI Shock" will likely be far less disruptive than the "China Shock" because it's hitting fundamentally different workers in fundamentally different places. The researchers found that the people and communities most vulnerable to AI displacement are actually the ones best equipped to adapt.
The data shows stark contrasts between the two groups of exposed workers. Only 5 percent of American workers faced serious exposure to the China Shock — roughly 6 million people, all in manufacturing industries where Chinese imports rose by 15 percentage points or more as a share of industry output between 1991 and 2007. In footwear, imports jumped 77 points; in toys and sporting goods, 70 points; in computers, 54 points. The majority of China-shocked workers had a high school degree or less, with only 6 percent holding advanced degrees. AI exposure tells a different story. Depending on how it's measured — from workers whose jobs are 50 percent exposed to AI task automation up to those 90 percent exposed — between 1 and 27 percent of the workforce faces potential disruption. But the majority of AI-exposed workers hold bachelor's degrees or higher, and those most exposed are three times more likely to have advanced degrees than just high school diplomas. China-shocked workers were evenly spread across wage levels, but AI-exposed workers cluster heavily in the top third of earners, with nearly two-thirds of the most exposed group making top-tercile incomes.
The report finds that education and earnings dramatically affect workers' ability to bounce back from job loss. "Low wage manufacturing workers in highly exposed industries lost about 1.2 years of initial earnings over the following 16 years," the authors write, while high-wage workers "were able to leave exposed firms and industries without large earnings losses." According to the researchers, this pattern held across multiple economic shocks beyond China, including the Great Recession and the decline of coal mining. Census data spanning over 80 years shows unemployment for the least educated third of workers runs at least double that of the most educated third.
Geography matters as much as demographics. The China Shock was geographically concentrated in manufacturing towns in the Rust Belt and South — the 10 hardest-hit commuting zones faced 5.8 times the average exposure, and the next 40 faced 3.2 times average exposure. These places also had below-average education levels, making adaptation harder. AI exposure is far more evenly dispersed, with the top 10 commuting zones only 1.2 times the average. The most AI-exposed places have above-average shares of college-educated residents, the opposite pattern from China-shocked communities. The report explains that when job losses are scattered nationwide rather than concentrated in a few towns, they blend into normal economic turnover — roughly 30 million jobs are lost every year to layoffs and closures even outside recessions, yet unemployment stays low because workers reallocate. Concentrated shocks overwhelm local labor markets and create demand shortfalls that monetary policy can't easily fix.
The authors acknowledge two worst-case scenarios: AI could displace workers faster than the China Shock did, or it could hit a much larger share of the workforce than expected. But they find both unlikely. Rolling out AI at massive scale requires computing hardware that takes time to build, and data center construction already faces constraints and policy pushback. More importantly, if the AI Shock does prove enormous and widespread, its very size would make economic policy more effective at offsetting demand effects, unlike the localized China Shock where monetary policy was mostly useless. The bottom line: "The AI Shock is aimed at the kinds of people and places that actually weathered the China Shock just fine."