Half of young workers feel guilty when they use AI to do their jobs. Yet the same skills are fast becoming something their employers demand. That is the bind facing Gen Z at work, according to a new global survey from the employment platform Employment Hero.
The company calls it “the AI paradox.” In the UK, 41% of workers say using AI to produce work makes them feel guilty, a figure that climbs to 51% among Gen Z. Across the four markets it polled, more than four in ten AI users say using the tools feels like cheating.
The data comes from a vendor with its own interest in the story, so treat it as a signal rather than gospel. Employment Hero surveyed 8,744 leaders and employees across the UK, Australia, Canada, and New Zealand.
A dirty little secret #
The guilt has a cost, because people tend to hide what makes them uneasy. Ria Kaur, a university student moving through internships, put the feeling plainly. “AI can feel like my dirty little secret,” she said. She blames a stigma that paints young people as lazy. Use AI at work, she said, and people assume you are cutting corners, even when you are using it to understand a task or prepare for a meeting.
The survey suggests bosses have not clocked the unease. Some 60% of employers think their staff view AI positively. Yet more than four in ten of those staff say using it feels like cheating.
The view from the boardroom and the view from the front desk simply differ. Tellingly, while 65% of employers say AI is speeding up the business, 63% of employees say it has created more work, in the form of checking the AI’s output.
The skill that now beats a degree #
Here is the paradox. The very thing that makes them uneasy is the thing bosses increasingly screen for. AI skills already outrank prior work experience, and even a university degree, on employers’ wishlists.
Across all firms they rank sixth, at 31%, behind old staples like work ethic (56%) and communication (46%). But among the most AI-focused firms, AI skills rank first.
The demand is showing up in the fine print of job ads. Employment Hero says mentions of AI skills are up 235% in a year, and that “Claude” is the fastest-growing skill term of 2026. Nearly six in ten businesses have already reweighted their hiring criteria towards AI, and another quarter plan to.
The workforce has not caught up. Six in ten employees rate their own AI ability as low or average. So young workers are teaching themselves. More than half said they learned AI on social media, rising to 56% in the UK.
Under-25s, the group entering a job market that now values AI over a degree, are 1.8 times more likely than the over-55s to say the training is their own job, not their employer’s.
The bar keeps moving #
The ground is also shifting under them. Even the definition of “AI” keeps moving. As Fortune reports, the billionaire Citadel founder Ken Griffin argues that most of what companies call AI is really older machine learning in a fashionable label.
One specific case changed his mind. A member of his own staff built an agentic system to reproduce and check academic finance papers. That job normally takes an expert six to eight weeks, Griffin said. The system did it in two to three hours per paper.
“This is not just a white-collar job,” he said. “This is a master’s or PhD-level job.”
His conclusion is double-edged for Gen Z. Griffin predicts a “golden age of entrepreneurial activity,” where small teams running AI agents can take on incumbents that once needed dozens of staff. That is a rare opening for young founders. It also means the skill bar for everyone else keeps climbing, even as the label on that skill stays slippery.
A new kind of school #
The education system is starting to react. On the same day the guilt survey landed, a two-year college for Gen Z, called Campus, launched a cybersecurity course for the AI era.
The pitch is a direct answer to the skills panic. More than a quarter of a million cybersecurity jobs sit unfilled, the college says, while AI makes attacks faster and cheaper. Its students take a dedicated “AI in Cybersecurity” module and train for industry certificates. The college counts Sam Altman and Shaquille O’Neal among its investors. It is betting on demand for what it calls AI-native talent.
It is one small example of a wider scramble. Employers want AI skills, formal training is thin, and so young people learn where they can, on social media, in side projects, and now in courses for the moment.
Why it matters #
Employment Hero’s own reading is upbeat. Its report, fronted by chief executive Ben Thompson, argues that AI is acting as a job creator, not a job killer: firms with AI at their core grew their entry-level headcount over the past two years at nearly twice the rate of non-adopters. The anxiety, in this telling, is a growing pain.
But the guilt is real, and it too has a cost. If people use AI in the shadows, managers cannot see its true impact, manage the risks, or spread good habits. The fix is not complicated, even if it is not easy. Give clear guidance, spell out what counts as fair use, and let a generation use the tools in the open.
As Employment Hero’s UK managing director Kevin Fitzgerald put it, AI at work should not feel like cheating. It should feel like using any other tool that helps you do the job.
Get the TNW newsletter #
Get the most important tech news in your inbox each week.