Your Employees Are Using AI. Are You Training Them or Losing Them?
Sign Up For Daily Job Market News
ZipRecruiter’s 2026 AI Worker Report found that 82.5% of U.S. workers use AI on the job to some degree, and 78% use it at home. Most of them like it. 52.6% say AI made their work at least somewhat more enjoyable, and only 5.5% say it made work less enjoyable.
Rank changes the answer. 76.5% of senior leaders say AI improved their job satisfaction, compared with 39.8% of entry-level workers. Training follows the same pattern. 62.8% of all workers got some AI training from their employer, but only 28.9% got extensive training. 83.5% of senior leaders got training, compared with 45.3% of entry-level workers. Overall, 37.2% of workers are teaching themselves, and that jumps to 54.7% for entry-level workers.
The pressure is real. 49.9% say they’re expected to produce more, and 44.6% say AI is making their technical skills outdated faster. Among daily AI users, that figure is 67.3%.
Employers are hiring for it, too. Job postings that call for using AI grew 159% in a year, according to ZipRecruiter, compared with 53% growth for jobs that develop or deploy AI. AI-enabled non-tech jobs now make up 4.3% of ZipRecruiter postings, up from 1.9%.
Workers don’t agree on what comes next. 68.1% say AI has changed or will change their career path. 25.1% plan to leave their jobs because of how their employer views or uses AI. For their next job, 30.5% want less AI involvement and 20.4% want more.
ZipRecruiter surveyed 1,500+ verified U.S. workers online from September 8 to 21. Labor economist Nicole Bachaud wrote that “as expectations rise, the workers without access to formalized training are the ones most at risk of falling behind.” Chief People Officer Amy Garefis added, “Formal training is only part of the picture,” and career expert Sam DeMase said, “Job seekers can’t wait for their next employer to provide them with AI training.”
The number I keep coming back to is the gap between the top and the bottom. Senior leaders got the training and they like the results. Entry-level workers mostly got neither. As I covered in September’s study of AI-adopting companies, junior employment fell 3% over 5 years at those firms, so the people with the least training are also the ones whose jobs are changing fastest.
Most of those junior workers are figuring it out on their own anyway, and if you’re early in your career, that’s your opening. When I’m talking to a candidate, “I taught myself and here’s what I built” beats “my company sent me to a course” every time. It also lines up with what employers were searching for on Upwork, where clients wanted specific, finished AI work over general tool familiarity.
Employers should look hard at the 25.1% who say they plan to leave. If you’re pushing people to use AI without showing them how, some of your best people will go somewhere that will. Stated intent is not the same as a resignation letter, and the survey can’t tell us how many will act. Even so, Monster’s survey found that 90% of workers say employers expect more loyalty than they give back, and an AI push with no training behind it adds to that feeling.
If you’re deciding where to start, give your entry-level team the same AI training your leaders got, and give them time to use it. Talk to us about hiring if you need people who already have those skills. And if you taught yourself, see what’s open right now.
270 Applicants Per Opening Means Your Best Candidate May Never Get Seen
More people are visiting job sites than a year ago. Similarweb counted 661.2 million visits to the top 100 U.S. job sites in September, up about 10% from 602.1 million a year earlier, according to Forbes. August was higher at 685.1 million, the most of any month this year, and Similarweb expects October to land 10.5% above last year. Indeed led September with more than 166 million visits, and LinkedIn visits were up more than 9%.
More traffic means more applications per job. Greenhouse CEO Daniel Chait says applications are up 20% from last year, with an average of 270 applicants for every open role. LinkedIn data shows applications per applicant up 30% from before the pandemic.
Job seekers are also less confident. In Pew’s survey from September 28 to October 4, 48% of Americans said they’re very concerned about people who want work not being able to find it, the highest reading Pew has recorded. It was 29% in October 2022. In Gallup’s May poll, only 28% said it’s a good time to find a quality job, down from 70% in June 2022.
Some of the postings people apply to aren’t real. The Wall Street Journal reported that about 19% of jobs posted on Greenhouse in the second quarter were fake.
Brad Hershbein, a senior economist at the W.E. Upjohn Institute, told Forbes that hiring is “slow, and it has been slowing over the last four-plus years.” J.T. O’Donnell, founder and CEO of Work It Daily, said “‘Spray and pray’ is probably where you’re seeing that big uptick in visits.” The experts Forbes quoted point job seekers toward fewer, better-targeted applications and networking.
The math explains why applying everywhere feels hopeless. If 270 people apply for one opening and about 1 in 5 postings may be fake, more applications won’t fix your odds. What works is what has always worked: get in front of a person. A referral, a recruiter who knows the hiring manager, or a direct note to the person who owns the role skips the pile completely.
Employers get no benefit from this either. When every opening gets hundreds of applicants, the person who would be great for the role gets buried along with everyone else, and your recruiters spend their week sorting instead of talking to people. It is one reason hiring managers bring in a staffing partner for roles they can’t afford to leave open. If you’re one of them, explore our staffing services.
Staffing your team doesn’t have to be hard.
Reach out and see how we can help.
Lean Teams Move Fast Until a Key Person Leaves
Dayforce surveyed 6,909 workers, managers, and executives at organizations with 100+ employees in Australia, Canada, Germany, New Zealand, the UK, and the U.S. The fieldwork ran July 27 to August 14, and the results are in its 17th Annual Pulse of Talent. Dayforce’s public release confirms the U.S. headline. 67% of U.S. respondents say AI adoption is outpacing workforce preparation. The full report puts the global average at 62%, making the U.S. the highest of the 6 countries.
Most results aren’t broken out by country, so the rest of the findings describe the full sample, which includes the U.S. Dayforce’s release reports that 51% say AI is reducing early-career opportunities, and 49% say AI now handles the tasks that used to teach people their jobs. 60% say understaffing has become normal, and 55% say losing several key people at once would cause significant or severe disruption. 73% say lost workforce capabilities are harder to rebuild than organizations expect.
Workers don’t expect AI savings to reach them. If AI saves time, 40% expect their employer to give them more work, and 39% expect layoffs or slower hiring. Only 26% expect the time to go toward better work. Leaders and workers also see training differently: 56% of executives say employees have time at work to build new skills, compared with 27% of workers.
Two more numbers come from the full report. 58% of workers are job hunting or open to the right offer, up from 48% in 2025, the first increase since 2022. And 32% of U.S. organizations land in the report’s high-risk group, the lowest of the 6 countries. Across all countries, the high-fragility share is 35%.
Dayforce Chief People Officer Amy Cappellanti-Wolf said: “Too often efficiency today comes at the expense of future resilience.” The report defines “experience debt” as “a future capability cost that accumulates when people lose opportunities to learn by doing.”
The U.S. results pull in two directions. Our companies look the most stable of the 6 countries, yet our workers are the most worried that AI is moving faster than their training. I think both are true. U.S. companies are moving fast on AI and operating lean, and so far it’s working. Workers can see that they aren’t getting the time or the entry-level experience to keep up.
The 58% is the number hiring managers should watch. Flight risk fell 3 years in a row and is back up 10 points. If your best people are the only ones who know how something works, this is when you find out what that costs. Write down who owns which process, get a second person trained on each one, and fix the entry-level learning gap before it becomes a skills gap.
A Government Set the Headcount Target Before Showing How AI Will Do the Work
Bloomberg reported today that the Dutch government plans to cut about 25,000 central government jobs over several years as part of a push to use AI to work more productively. The goal is to save €900 million a year by 2027, rising to €3.5 billion a year from 2035. The cabinet will present its AI plans early next year.
A national government just set a headcount target based on the savings it expects from AI, before showing how the AI will do the work. I’d expect plenty of employers to copy that approach. Barclays announced about 5,000 cuts last Thursday in a cost plan Reuters tied to AI, and the Dutch plan follows the same order, with the target first and the detail later.
If you’re an employer thinking about the same move, test the plan before you publish the number. Ask which tasks the AI will take over, who will review its work, and what happens to the learning that junior roles used to provide. Dayforce’s “experience debt” finding is the cost of skipping those questions.
