Tufts University releases the first American AI Jobs Risk Index

There is a certain irony at the center of a new analysis from Digital Planet at Tufts University's Fletcher School. The regions of the United States most deeply invested in developing artificial intelligence, Silicon Valley, Boston, Washington, Seattle, also face the highest projected risk of workforce displacement from the same technology they are building.

That geographic inversion is one of the more striking findings in a report that attempts something prior analyses have largely avoided: mapping not just which jobs AI could affect, but which workers are likely to actually lose them, where those workers live, and how much income is at stake.

The headline figure is substantial. About 9.3 million U.S. jobs could be displaced within the next two to five years. Depending on the speed of AI adoption, that range extends from 2.7 million at the low end to 19.5 million at the high end. The annual wages tied to those jobs sit between $200 billion and $1.5 trillion, with a midpoint estimate of roughly $757 billion.

The distribution of risk challenges a widespread assumption about automation. For decades, the dominant narrative has been that machines replace repetitive, physical, or predictable tasks while leaving knowledge workers largely untouched. This analysis points toward a different pattern.

Industry-wide, AI displacement vulnerability averages about 6 percent across the economy. But that average conceals enormous variation. Information sector jobs face an 18 percent risk. Finance and Insurance, as well as Professional, Scientific, and Technical Services, each sit at 16 percent. Meanwhile, occupations defined by physical labor in unpredictable environments, roofers, orderlies, dishwashers, face displacement risk under 1 percent.

At the occupational level, the exposure is sharper still. Writers and authors face a 57 percent displacement rate. Computer programmers and web designers each sit at 55 percent. Historians top the list at 67 percent of their tasks expected to be automatable.

"We already know that AI is not just automating routine tasks, it is moving up, targeting the cognitive and analytical work that defines high-skill, high-wage careers," said Bhaskar Chakravorti, dean of global business at The Fletcher School and chair of Digital Planet.

The unsettling implication, which the data make explicit, is that the safest jobs right now are often the lowest-paid ones. Approximately 38 percent of American workers face less than 1 percent displacement risk. They are not protected because their work is hard to automate in principle; they are protected because their work is physical, spatially embedded, or unpredictable in ways current AI systems handle poorly.

The analysis maps displacement risk down to individual metro areas and states, revealing a concentration that follows the geography of the knowledge economy.

New York, Los Angeles, Washington, Chicago, Dallas, San Francisco, 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. Washington, D.C. tops the state-level rankings at 11.3 percent, followed by Massachusetts, Virginia, Maryland, Washington state, and Colorado.

University-centered metros also rank near the top: Durham-Chapel Hill, Boulder, Ann Arbor, Ithaca, and Madison all appear among the 25 most vulnerable areas. The report describes these clusters as "Wired Belts," places where the concentration of technical talent and AI-adjacent work creates both innovation capacity and economic exposure.

The largest total job and income losses are projected for California, Texas, New York, Florida, and Illinois, not because these states are more vulnerable proportionally but because of their sheer economic size.

Not all risk is static. The analysis identifies 33 occupations, covering roughly 4.9 million workers, that the researchers label "tipping point" roles. In these jobs, displacement risk could jump from under 10 percent to over 40 percent depending on how quickly AI tools improve and spread through industries.

Even workers whose jobs center on AI itself are not exempt. More than one million people employed to study, build, or report on artificial intelligence face displacement rates between 26 and 55 percent, a dynamic the report does not attempt to fully explain but flags as significant.

A single quantitative relationship runs through much of the data: for every one percentage point increase in automation, the study projects a 0.75 percentage point loss in jobs. The relationship is not one-to-one, but it is close enough to suggest that efficiency gains from AI may translate fairly directly into workforce reductions rather than being absorbed through growth in output or the creation of new roles.

The analysis notes, with some candor, that it does not include estimates of job creation from AI due to limited available data. That omission shapes the picture considerably. Whether displaced workers find new roles in AI-adjacent fields, whether new industries emerge to absorb them, and on what timeline, are questions the index does not answer.

The geographic concentration of risk has a policy dimension that the report highlights directly.

States facing higher AI exposure are legislating on the technology at four times the rate of states with the least exposure. That pattern makes intuitive sense: Massachusetts, California, and Washington have both the most to lose economically and the administrative capacity to respond.

But a December 2025 executive order has complicated that dynamic, directing the Justice Department to challenge certain state-level AI laws and threatening to withhold federal broadband funding from states that proceed with their own frameworks. The resulting tension between state and federal approaches to AI governance could significantly influence how the labor market responds in the coming years, particularly in the regions most exposed.

Chakravorti framed the urgency plainly. "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," he said.

The report's approach differs from earlier displacement studies in a specific way worth noting. Rather than measuring which jobs AI could theoretically affect, it estimates the probability that exposure translates into actual job loss, then connects that estimate to income data and geography. That shift in methodology produces a more actionable picture, even if the projections carry wide uncertainty ranges.

For workers in high-exposure occupations, particularly those in technical, analytical, or creative fields in major metro areas, the findings suggest that the transition pressure may arrive faster and with less warning than conventional forecasts have implied. For policymakers, the geographic specificity of the risk map provides a basis for targeted intervention rather than broad responses to a diffuse problem.

For the technology sector itself, the analysis lands as something of a mirror. The regions building the infrastructure of the AI economy are also the ones most likely to experience its labor market consequences first, and the workers whose skills most resemble the capabilities AI is developing fastest are among the most exposed. That is not a reason to slow development, but it is a reason to think carefully about what policy, retraining, and transition support need to look like before the tipping points arrive.

Research findings are available online at the Digital Planet.

The original story "Tufts University releases the first American AI Jobs Risk Index" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.