Cognizant Study: 93% of U.S. Jobs Now Have Tasks Exposed to AI
The most striking finding: the current level of AI job exposure has already surpassed what the original study forecast for 2032. That means the workforce is facing disruption roughly six years ahead of schedule. Exposure scores across all occupations are running 30% higher than their projected 2032 levels, and the annual growth rate of AI exposure has jumped from 2% to 9% since the first assessment.
At the top of the exposure spectrum, financial managers lead all categories with 84% of their tasks now replicable by AI. Computer and mathematical roles follow at 67%, with business and financial operations, office and administrative support, and legal occupations all clustered in the 60% to 68% range. Management and C-suite positions sit at roughly 60% exposure, a particularly notable jump from the 25% originally projected for CEOs by 2032.
The study attributes the acceleration to three AI capability breakthroughs: multimodal systems that can process images and spatial data, expanded reasoning models capable of handling complex cognitive tasks, and agentic AI systems that can autonomously execute multi-step workflows across tools and platforms.
Jobs with the lowest AI exposure tend to involve physical presence, manual dexterity, or real-time situational judgment. Construction workers (12%), mechanics (17%), and installation and repair technicians (20%) remain among the least affected. However, even these categories are seeing measurable increases in exposure. Construction jumped from 4% to 12% since 2023, and transportation went from 6% to 25%.
Weekly Unemployment Claims Hold Near Historic Lows
The Department of Labor reported 212,000 initial jobless claims for the week ending February 21, up slightly from 208,000 the previous week. The four-week moving average held steady at 220,250. Continuing claims dropped to 1,833,000, down 31,000 from the prior week, with the insured unemployment rate unchanged at 1.2%.
One notable trend: continuing claims from former federal civilian employees rose to 12,657, a 66% increase from 7,612 a year ago, reflecting ongoing federal workforce reductions. No state reported a significant increase in initial claims. Twelve states reported decreases of more than 1,000, led by New York, Pennsylvania, and New Jersey, across construction, transportation, and healthcare sectors.
Chicago Fed Data Shows Hiring Rates Slipping for Unemployed Workers
The Federal Reserve Bank of Chicago released its advance labor market indicators for February, forecasting an unemployment rate of 4.28%, unchanged from January but up from 4.17% a year ago. The layoffs and separations rate held flat at 2.08%, confirming that employers are not cutting workers at elevated rates.
The more telling figure is the hiring rate for unemployed workers, which dropped to 45.23% from 47.49% a year ago. That means unemployed job seekers have a measurably harder time finding work than they did twelve months ago. The model gives 41% odds that the unemployment rate decreases, 28% odds that it stays flat, and 30% odds that it increases.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Frequently Asked Questions
According to Cognizant’s “New Work, New World 2026” study, 93% of U.S. jobs have at least some tasks that AI can currently perform. This does not mean 93% of jobs will disappear, but it does mean that the vast majority of occupations include work that AI systems can handle today.
Financial managers have the highest AI exposure at 84% of tasks. Computer and mathematical roles follow at 67%. Business and financial operations, office and administrative support, legal occupations, and management roles all fall between 60% and 68%. These are primarily knowledge work positions that involve reading, writing, analyzing, and communicating.
Jobs requiring physical presence and manual skills remain the least exposed. Construction trades sit at 12%, mechanics at 17%, and installation and repair technicians at 20%. Protective services and personal care roles range from 20% to 29%. However, even these categories have seen increases in AI exposure since 2023.
Cognizant estimates $4.5 trillion in U.S. labor value could theoretically shift from human workers to AI systems. Globally, that figure approaches $15 trillion. These numbers reflect the current state of AI capability, not projections about future improvements.
For the week ending February 21, 2026, initial jobless claims totaled 212,000. The four-week moving average was 220,250. Continuing claims dropped to 1,833,000. These figures indicate that layoff activity remains historically low.
Chicago Fed data suggests yes. The hiring rate for unemployed workers fell to 45.23% in February 2026, down from 47.49% a year earlier. While the overall unemployment rate has only edged up slightly to 4.28%, the declining hiring rate means job seekers face longer searches and fewer opportunities than they did twelve months ago.
Cognizant’s study identifies three capability advances: multimodal AI (systems that process images, video, and spatial data), expanded reasoning models (systems that handle complex cognitive tasks), and agentic AI (systems that autonomously execute multi-step workflows). Together, these have pushed AI exposure scores 30% above what was originally projected for 2032.
While overall layoff numbers remain low, continuing unemployment claims from former federal civilian employees rose to 12,657, a 66% increase from 7,612 a year ago. This trend reflects ongoing federal workforce reductions and restructuring.
View Full Transcript
0:00 — Cognizant: 93% of U.S. jobs now exposed to AI
Today’s job market headlines include a new report that says the labor market is currently stuck in neutral. You’re surprised to hear that? You probably you probably are. And that unemployed workers are having a harder time finding jobs than they did a year ago. And also this morning’s weekly unemployment claims came in low again, which sounds good on the surface, but I’ll get into those details in a minute because the big story today is a new report from Cognizet. They’re one of the largest IT services and consulting firms in the world. They studied more than 18,000 tasks across a thousand jobs to figure out where AI can actually do the work.
Now they originally published research similar to this in 2023, and they predicted it would take until 2032 to reach the level of disruption that we’re currently experiencing. So, in other words, six years ahead of schedule from what they originally predicted. And I think that headline alone is so indicative of what AI represents, what we see. It’s evolving so rapidly. It is truly unlike anything we’ve experienced before, and its effect on the job market is going to be profound. So here’s the headline number from the report. So not in five or six years, right now. And all that adds up to about 4.5 trillion in US labor value that could shift from human workers to AI.
And globally, that number reaches around 15 trillion. So the speed is picking up. That’s what we’re seeing. And as I said, when they first measured this in 2023, AI exposure was growing at about 2% a year. Now it’s growing at 9% a year. And I expect that will increase. As you, if you listen to me at all, you know that that is my regular stance on this, and everything I see only supports that getting stronger. So this is not a gradual shift. This is a drastic acceleration. That is what is happening. Now, to clarify, exposed, so these jobs are exposed to AI. It doesn’t mean that 93% of the jobs will disappear. It just means that those jobs have at least some tasks that AI can handle today.
So some jobs have a lot of those tasks, some have few. And so the question that should be on everyone’s mind is where your job falls on that spectrum. And that’s what this report breaks down. So I’ll share the highlights, the ones, or perhaps the lowlights, depending on your perspective. At the top of the list of exposed jobs are financial managers. 84% of their tasks can be done by AI. That’s the highest of any job category in the study. And if you think about what financial managers actually do every day, they analyze reports, they build forecasts, review budgets, track performance. Those are the things that AI is already very good at and continually getting better at.
So it doesn’t mean you don’t need a financial manager. But one financial manager with AI that’s powered by AI can probably do the work that used to take two, three, maybe more. And that’s the trend that I see continuing the most. It’s not that the jobs are going to go away, it’s that everyone who’s leveraging AI will be infinitely more efficient, more productive with it, able to do more. And so that the numbers, while no one really knows how that’s going to shake out, but that that’s what concerns me is that we’re going to need fewer people to do the same level of work. Also, computer and mathematical roles, 67% of those tasks can be replaced by AI. And that shouldn’t be a surprise if you’ve been paying attention. Software development is essentially ground zero for AI disruption right now.
Linus Torvald, who created Linux, he says that AI now codes better than him. I mean, if that is a profound statement by someone who is one of the best who’s ever lived. He’s saying that. And Claude Code, if you if you’re following Claude at all, they’ve recently said the company behind Claude Anthropic has said that the latest version of Claude was written by Claude. So that is what is taking place here. And other job categories, business and financial operations are high on the list, office and admin support and legal jobs are also at risk. So that those shouldn’t be a surprise if you’ve been paying attention. I mean, those jobs uh come up on almost every report regarding AI exposure, but here’s one that everyone will find interesting.
Management and C-suite jobs are at 60%. They’re also in the top uh exposed jobs. And then so that management number is really worth pausing on because in the 2023 study, CEOs had a projected exposure score of just 25% by 2032. And today it’s already over 60. So the people making decisions about whether to adopt AI have more than 60% of their own tasks exposed to it. So that irony shouldn’t be lost on anyone. And it that is consistent with what we saw in our Q1 survey that we did where the executives who claim to use AI more than staff-level employees, executives seem to be using AI the most, they’re also the most afraid of it. And I think that is a consistent theme as well. The people who use it the most realize how impactful it is, how powerful it is, and as a result, they’re the most fearful of it.
So I found that really interesting. And a lot of people always say, well, what about the people making the decisions, you know, who are displacing workers? They’re gonna be displaced too, to a significant degree. Again, we’re gonna need fewer people in pretty much across the board in white-collar jobs. Now, let’s look at the other end of the spectrum, and that is the jobs that uh are going to be least affected or at least affected right now. Construction. Only 12% of construction tasks can be replaced by AI right now. Mechanics, installation and repair uh professions, uh professionals, protective services. So we need physical, we need people to be there physically, of course, until the robots take over.
So for now, those jobs are among the safest. If you if it requires dexterity, if it requires human presence, right, then those jobs are safe. And healthcare support was also on that list. So healthcare, anything physical right now, that are this, those are the least exposed to AI. It shouldn’t be a surprise. What what the the big takeaway from this is just how fast this has evolved. Now, interestingly, construction exposure did jump from 4% in 2023 to 12% today. Now, the cognizant CTO said that the surprise was that some of the tasks we were expecting to be automated later on are already being automated. The AI impact in physical jobs isn’t just about robots swinging hammers, although I do believe it eventually will be. If you’ve seen some of the new robots that are coming out, but it’s about AI handling uh the inspection, the planning, the documentation, the scheduling. So anything that supports the hands-on work, that is where the AI is going to continue to creep in.
The bottom line is four and a half trillion in labor value that companies can potentially shift to AI. They’re going to do it. And believe me, I don’t like it. I own a staffing company. This is not good for my business, to say the least. We’re going to be affected, but we’re going to be okay. But the generations behind us, it’s it is the Wild West, it’s a brand new world, and I’m concerned. So that’s why one of the reasons I share this as it comes out.
8:04 — Weekly unemployment claims hold near historic lows
Also, today the weekly initial jobless claims came in at 212,000 for the weekending February 21st, and that’s up 4,000 from the previous week. So not a big shift, but it is a low number. A year ago it was 243,000, so that’s a positive trend. We don’t see a lot of positive trends in the job market lately. That is one. The four-week moving average came in at 220,000,250, so that really didn’t move. Continuing claims, and those are the number of people actively collecting unemployment benefits. That dropped uh to 1.83, which is down 31,000 from the prior week.
We’re not seeing big cuts, we’re not seeing great hires, low hire, low fire environment. That’s what’s happening right now. Um, also at the state level, no state reported a significant increase in initial claims last week. So nothing really to share there. Um so the bottom line for right now is that if you’re employed, you’re in decent shape. That’s kind of what these these numbers tell us. But we’ll see. We’ll see what happens in the in the weeks and months ahead.
9:15 — Chicago Fed: Hiring rates slipping for the unemployed
Also, today the Federal Reserve Bank of Chicago released its advanced labor market indicators for February. That came out this morning. Um, and this is a data that puts the weekly claims into some context. The Chicago Fed tracks two things that matter most. One is that they they look at how many people are flowing into unemployment, meaning they’ve been laid off or let go. And then, second, how many unemployed people are flowing back out, meaning they’ve found new jobs. So this really puts some clarity around just the general unemployment claims. And what they forecast for February is 4.28% unemployment, which is effectively unchanged from January.
A year ago it was 4.17, so unemployment is about a tenth of a point higher than it was a year ago, which isn’t dramatic, but although it is shifting in the wrong direction, the hiring rate for unemployed workers went down a couple points. So if you were unemployed in February of last year, you had an easier time finding a job than you do today. It’s not a huge percentage difference, just two points. It’s meaningful to those involved. Don’t get me wrong, but that’s where we are right now, as I as I continue to say, unemployed people aren’t getting hired at a great rate, it’s getting weaker, but we’re not seeing huge cuts. So now that is the forecast for February. We’ll have the final numbers pretty soon. So that’s it for today. Check out the cognizant study. We’ll link it, of course. Really, really interesting. It’s a big one, it’s meaningful to see how much they shifted in just a couple of years.
10:51 — Today’s fun fact
But before we go, here’s the fun fact for today. Apparently, brainstorming isn’t as beneficial as as I thought. Maybe you thought too. Research shows people are actually less creative in groups than when working alone. So no more brainstorming. Do your thinking alone, everyone. Thanks for listening. Please like, subscribe, share with anyone who you think might be interested, and I will look forward to talking to you tomorrow.
