The regions that built artificial intelligence are the regions most likely to be dismantled by it.

That is the central finding of the American AI Jobs Risk Index, released this week by Digital Planet at Tufts University’s Fletcher School — the first comprehensive effort to map not just which jobs AI could theoretically touch, but which workers are likely to actually lose them, where they live, and what it will cost.

The numbers are stark. Approximately 9.3 million US jobs face displacement over the next two to five years under the index’s median adoption scenario, with a plausible range of 2.7 million to 19.5 million depending on how quickly AI tools spread through the economy. The annual wages tied to those jobs sit at a midpoint estimate of $757 billion — roughly the GDP of Belgium. Under faster adoption scenarios, losses approach the economy of South Korea.

The Wired Belts

The report introduces a piece of framing worth paying attention to: the “Wired Belts.” Just as deindustrialization hollowed out the Rust Belt, the researchers argue, AI-driven disruption is poised to hit the knowledge economy’s own heartlands first — and hardest.

Major urban centers — New York, Los Angeles, Washington, San Francisco, Chicago, Dallas, and Boston — each face at least $20 billion in projected annual income losses. The San Jose metro area, home to Silicon Valley, leads the country in proportional job risk at 9.9 percent. University towns including Durham-Chapel Hill, Boulder, Ann Arbor, Ithaca, and Madison rank among the top 25 most vulnerable metros.

These are the places with the highest concentrations of technical talent, analytical work, and AI-adjacent industries. They are also precisely the capabilities AI is now learning to replicate.

Follow the Money Down

At the industry level, vulnerability averages roughly 6 percent across the economy. That average conceals enormous variation. Information sector jobs face an 18 percent displacement risk. Finance and insurance, along with professional, scientific, and technical services, each sit at 16 percent.

At the occupational level, the picture sharpens. Writers and authors face a 57 percent displacement rate. Computer programmers and web and digital interface designers each sit at 55 percent. Historians top the list at 67 percent of tasks expected to be automatable. The greatest total income losses fall not on niche roles but on the big employment categories: software developers, management analysts, and market research analysts — reflecting both high salaries and large worker populations.

Half of all projected job losses come from just 26 occupations. Eight occupations drive a quarter of total displacement. The report identifies 33 “tipping point” occupations covering 4.9 million workers — roles where displacement risk could swing from under 10 percent to over 40 percent depending on the pace of AI adoption.

The Safe Zone Is the Poverty Zone

Approximately 38 percent of American workers are effectively AI-proof, facing less than 1 percent displacement risk. These are the country’s lowest-paid jobs — roofers, orderlies, dishwashers. Physical, manual labor in unpredictable environments. The occupations AI cannot touch are largely those the economy has always undervalued.

The relationship between augmentation and displacement makes for equally uncomfortable reading. The index finds that for every one percentage point increase in automation, it projects a 0.75 percentage point job loss. The more AI helps you do your job, the more expendable you become. Finance professionals, teachers, creative workers, accountants, and legal professionals are next in line.

The Political Collision

The geographic concentration of risk carries direct political consequences. States most exposed to AI job displacement — DC, Massachusetts, Virginia, Maryland, Washington, and Colorado — are legislating on the technology at four times the rate of the least-exposed states. A rational response from places that have both the most to lose and the administrative capacity to respond.

But a December 2025 executive order has complicated that dynamic. The federal government directed the Justice Department to challenge certain state-level AI laws and threatened to withhold federal broadband funding from states that proceed with their own regulatory frameworks. The states with the most at stake are being told to stand down.

DC itself presents a particular irony: it carries the highest vulnerability of any US jurisdiction at 11.3 percent and has produced zero AI legislation.

“Our index makes clear that the question is no longer whether AI will displace significant numbers of workers, but in which states and cities, how fast, and whether we are prepared by taking pre-emptive action,” said Bhaskar Chakravorti, dean of global business at The Fletcher School and chair of Digital Planet.

What the Index Does — and Doesn’t — Measure

The American AI Jobs Risk Index covers 784 occupations, 530 metro and non-metro areas, 50 states, and 20 industry sectors, drawing on 15 years of labor market data and current AI adoption research. Its key methodological advance is measuring vulnerability — the probability that AI exposure translates into actual job loss — rather than mere exposure, which is where most prior studies from Goldman Sachs, MIT, Yale, and Stanford have stopped.

The index deliberately excludes job creation effects, citing a lack of robust data. The researchers note they intend to incorporate those dynamics in future updates. That omission is honest but shapes the picture considerably. Whether displaced workers find new roles, whether new industries emerge to absorb them, and on what timeline — those questions remain open.

As an AI newsroom reporting on research about AI’s labor market impact, we have a stake in this story and no intention of pretending otherwise. The finding that over one million workers whose jobs involve studying, building, or reporting on AI face displacement rates between 26 and 55 percent is, to put it mildly, close to home.

Sources