Tufts Just Mapped AI Job Risk Across Every City and State in America & the Results Are Counterintuitive
The median projection: 9.3 million U.S. jobs are vulnerable to AI-driven displacement within the next two to five years. The range runs from 2.7 million under slower adoption to 19.5 million if agentic AI adoption accelerates faster than expected. The income at stake is staggering, between $200 billion and $1.5 trillion in total annual wage losses, representing somewhere between 2% and 15% of all U.S. wage income.
The single most important number in the entire report may be this: for every one percentage point increase in job automation, researchers found a 0.75 percentage point actual job loss. That ratio effectively dismantles the argument that AI simply moves people to higher-value work. At scale, that conversion rate is devastating. The report’s authors note that this wipeout, under the median scenario, is equivalent to wiping out the entire economy of Belgium. Under faster adoption, it approaches the economy of South Korea.
The industry breakdown puts average vulnerability across all sectors at about 6%, but three industries are well above that floor. Information comes in at 18%, Finance and Insurance at 16%, and Professional, Scientific, and Technical Services at 16%. Across all vulnerable occupations combined, $757 billion in annual income is at risk, representing 7.5% of total U.S. labor income. The most exposed individual occupations are Writers and Authors (57%), Computer Programmers (55%), and Web and Digital Interface Designers (55%).
Here’s the counterintuitive part. The places most at risk aren’t struggling towns. They’re innovation capitals, the places that were supposed to be future-proof. New York, Los Angeles, Washington D.C., San Francisco, Chicago, Dallas, and Boston each face at least $20 billion in projected annual income losses. The concentration extends to university towns and research hubs too: Durham-Chapel Hill, Boulder, Ithaca, Madison, and Ann Arbor all rank among the top 25 most vulnerable metros by percentage of local employment at risk. Among the top 25 major metros, Orlando, San Antonio, and St. Louis are among the least affected. At the state level, California, Texas, New York, Florida, and Illinois account for 40% of all projected AI job losses nationally.
One more finding worth sitting with: the 4.9 million workers the researchers identify as “tipping point” workers, people in 33 occupations that swing from under 10% to over 40% displacement within that same two-to-five-year window. Among them are more than a million people whose actual job is to study, build, or report on AI, facing 26–55% displacement risk. And the researchers flag something we’ve been tracking across multiple data sources: augmentation and displacement go hand in hand. The more AI helps you do your job, the more expendable you can become. Finance professionals, teachers, creative professionals, accountants, and legal professionals are all listed as next in line.
There’s also a political dimension. States most at risk of AI job losses are legislating four times more aggressively on AI than the safest states. The federal government recently directed those states to pull back on what it called “onerous AI regulation.” The report is careful not to take political sides, but the tension it identifies is real: the places with the most at stake are being told to stand down from the policies they’re pursuing to address it.
89% of 2026 College Graduates Are Worried AI Will Eliminate Entry-Level Jobs. Last Year It Was 64%.
Monster’s 2026 State of the Graduate Report puts the anxiety of the graduating class into hard numbers, and the trajectory is alarming. 89% of 2026 graduates now worry that AI could replace entry-level roles. Last year, that figure was 64%. A 25-percentage-point jump in 12 months isn’t noise. It’s a workforce reading the room.
The anxiety is shaping decisions at every level. 67% of 2026 grads say they’d accept a lower-paying job if it offered greater long-term security. Job security now ranks at 52% as a decision driver, overtaking career growth at 49% for the first time. These are young people who are supposed to be focused on building careers; instead, they’re focused on not losing a foothold before they’ve established one. 76% say the economy’s impact on their job prospects is a top concern.
Salary still matters, 68% rank it as a top factor, but the framing has shifted. 75% say they’d accept a job they expect to leave within a year if it provides immediate income. Nearly seven in ten say they’re more willing to compromise on their ideal role than they were a year ago. These aren’t irrational responses. Entry-level hiring has been contracting across white-collar fields for months, and these graduates are arriving to a market that has been quietly eliminating the roles that used to be their starting point.
On timeline expectations, 79% believe they’ll land a job within three months. But 35% expect their search to take four months or longer, and 15% anticipate it stretching past six months. That gap between optimism and realism is itself telling, and it mirrors the split we’ve been seeing in broader job seeker confidence data. People are holding onto hope while also hedging against a market that isn’t cooperating.
What can employers actually do with this? The graduates entering the workforce right now are motivated, adaptable, and, given all of the above, acutely aware that they need to prove their value quickly. Demonstrating stability where it genuinely exists, articulating clear career paths, and investing in AI training for new hires isn’t just good HR. It’s a competitive advantage in a market where young talent is watching closely for signals about which employers are worth committing to. If you’re looking to hire into this cohort or connect with the right employer as a new grad, the data suggests moving sooner rather than later.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Frequently Asked Questions
By percentage of local employment at risk, San Jose, Washington D.C., Boulder, Durham-Chapel Hill, and San Francisco lead the list. By absolute job losses, New York, Los Angeles, Washington D.C., Chicago, and Dallas top the rankings. At the state level, DC, Massachusetts, Virginia, Maryland, Washington, and Colorado are the most vulnerable by share of jobs at risk, while California, Texas, New York, Florida, and Illinois face the largest absolute losses.
Deliberately not. The researchers explicitly excluded job creation from this version of the index because they say the empirical evidence on net job creation remains too limited and speculative to include responsibly. The index focuses on vulnerability to displacement. They plan to incorporate job creation data in future updates as more reliable evidence becomes available.
Because AI’s current capabilities are strongest in the kinds of tasks that define knowledge work: writing, coding, analysis, research, and documentation. These are precisely the skills concentrated in university towns and tech metros. Physical, variable, and safety-critical work is far harder to automate, which is why manufacturing, construction, and agriculture show much lower displacement rates.
The data points in a consistent direction: build AI fluency, prioritize roles where you can develop hands-on experience quickly, and don’t hold out for the ideal role at the expense of getting started. Monster’s survey shows that nearly 70% of graduates are already willing to compromise on their ideal role, which is well calibrated to current conditions. What makes the biggest long-term difference is whether you’re building real skills and demonstrable output from day one.
View Full Transcript
Tufts University’s American AI Jobs Risk Index: 9.3 million jobs at risk & the cities that didn’t see it coming
Welcome back to Cornering the Job Market. Today’s workforce news and headlines include a new survey indicating that 89% of 2026 college grads are worried AI will replace entry-level jobs. We’ll get into all of that. Plus, this is the biggest week of the month for jobs data. I’ll share what’s coming and exactly what to watch out for. But first, a major news study from Tufts University just mapped AI job risk across every metro area and every state in the country. The study’s called the American AI Jobs Risk Index, and it projects that 9.3 million U.S. jobs are vulnerable to AI-driven displacement within the next two to five years.
That’s the median scenario. The range runs from$2.7 million on the low end up to 19.5 million under faster adoption. The income at stake is enormous. They’re projecting between$200 billion and$1.5 trillion in total wage losses, depending on how quickly companies move, and that accounts for between 2.5 to 15% of all U.S. wage income. Here’s one of the most important numbers in the whole report. For every one percentage point increase in job automation, the researchers found a 0.75 percentage point actual job loss. That should end any AI just moves people to higher value work arguments, because that conversion rate, even if it’s just close, is way too high because at scale, that math is devastating.
To put it in perspective, the baseline job loss projection is equivalent to wiping out the entire economy of Belgium. And under faster adoption scenarios, including the acceleration of agentic AI, it approaches the economy of South Korea. Now for the industry breakdown, average vulnerability across all sectors is about 6%. But three industries are way above that. Information at 18%, finance and insurance at 16%, and professional, scientific, and technical services were also at 16%. Across all vulnerable occupations,$757 billion in annual income is at risk, which is 7.5% of total labor income. But gets even more interesting when you look at the places that are most at risk. They aren’t struggling towns, they’re the areas that are considered innovation capitals. New York, Los Angeles, DC, San Francisco, Chicago, Dallas, and Boston each face at least$20 billion in annual income losses.
And it goes beyond the mega metros. University towns and innovation hubs are in the crosshairs too. Durham, Chapel Hill, Boulder, Ithaca, Madison, Ann Arbor. Among the top 25 major metros, Orlando, Santonio, and St. Louis are among the least affected. So geography here is counterintuitive, at least from what I would expect. Now at the state level, California, Texas, New York, Florida, and Illinois account for 40% of all projected AI job losses nationally. The least affected for absolute job losses, not a surprise, Wyoming, Alaska, Vermont, and the Dakotas. But the study identified 4.9 million tipping point workers across 33 occupations. These are people in roles that swing from under 10% to over 40% displacement within that two to five year window. Among them are more than a million workers whose actual job it is to study, build, or report on AI. They face between 26 and 55% displacement.
Now, one of the other findings that’s worth paying attention to is that augmentation and displacement go hand in hand. The more AI helps you do your job, the more expendable you become. So that includes finance professionals, teachers and professors, creative professionals, accountants, legal professionals. They’re all next in line for potential displacement. And the 38% of American workers whose jobs are considered AI proof, they hold the lowest paid jobs in the economy. So the safe zone is in the near poverty zone. That’s not good at all. There’s a political dimension here too. Of course there is. States most at risk of AI job losses are already legislating four times more on AI than the safest states. And the federal government recently directed those states to pull back on what it called onerous AI regulation. What a mess. You’d think this would be getting more press.
This is a very alarming study, but it’s just not. And we continue to hear from the very top at the federal level that there’s nothing to be concerned about with AI job displacement. When you see research like this, you just can’t believe that. Because what I find most concerning and something I just hadn’t heard before is that 0.75 conversion rate for every percentage point of automation, three-quarters of a percentage point in actual job loss is what it will amount to. When you run that math, it is truly devast devastating at scale. So that’s what we’re looking at here. 9.3 million jobs at risk concentrated in the cities and industries that were supposed to be future-proof. It doesn’t look like that’s the case.
Monster’s 2026 State of the Graduate Report: 89% worry AI will replace entry-level jobs
So let’s move on from there and let’s look at how people entering the workforce feel about all of this. Not good, right? Monster just released its 2026 State of the Graduate Report, which is a survey of recent and upcoming college graduates. The headline number is that 67% would accept a lower-paying job if it offered greater long-term security. But the standout data point is the AI anxiety jump. 89% of 26 grads now worry AI could replace entry-level roles. Last year, that number was 64%. Still unacceptably high, but that 25% point spike in 12 months is just awful to see. What a bad state of things for our young people.
When you pair that with the 76% who are concerned about the economy’s impact on their job prospects, we have a graduating class that is having to make decisions fundamentally different than any of those before them. Salary still matters, of course. 68% rank it as a top factor when evaluating job offers, but job security at 52% now outranks career growth at 49% as a decision driver. So again, these young people are having to focus on the entirely wrong things as they’re starting their career. But we collectively have put them in that situation because they’re reading the room correctly. They’re not looking ahead, they’re focused on today.
75% would accept a job they expect to leave within a year if it provides immediate income. Nearly seven in ten say they’re more willing to compromise on their ideal role than they were a year ago. On timeline, 79% believe they’ll land a job within three months, but more than a third, 35%, expect their search to take four months or longer, and 15% anticipate it stretching past six months. So while no one individually can change what’s happening to the economy overall, and you can’t change what’s happening at scale, we can make changes at the individual level. So if you’re an employer, do your best to demonstrate stability if it exists and articulate career paths where they exist for your new hires because there’s a lot of motivated young people right now.
I’m surrounded by them and my personal life as well as my professional one. So while we can’t change the bigger picture, we can make improvements one at a time. So take care of our young professionals. They truly are our future and they need help right now. That’s it for today.
The week ahead: JOLTS, ADP, Challenger, BLS, here’s what to watch
But before we close, let’s look at the week ahead because it’s gonna be a big one. Tomorrow’s a Jolt’s report. Tuesday is the ADP employment report. There are expected to be 37,000 private sector jobs added in March. That’s soft to begin with. We’ll see where we come in, maybe, maybe above, maybe we get lucky. I doubt it. I wouldn’t be surprised if it’s lower. Wednesday, Challenger will announce job cuts and we’ll also have the weekly jobless claims. And Thursday’s the big one. That’s the BLS jobs report. So we’ll see what is going on there. I will share all of it. I look forward to doing that as the week goes on. Maybe we’ll get some good news. Maybe we’ll get lucky. Who knows? We can always hope.
Fun fact: “Eating the frog” means doing your hardest task first
But here’s a fun fact before we go. If you’ve heard the term eating the frog, you know what that means? It means doing your hardest task first thing in the morning. Hopefully that hardest task is not going to be reading these jobs reports. Maybe again we’ll get lucky. Fingers crossed. Thank you for listening today. Please like, subscribe, share with anyone who you think might be interested. I’d appreciate that. Maybe someone else would too. And I will look forward to talking to you tomorrow.
