Gen Z Judges Colleagues Who Use AI at Work: What the Survey Really Found
Key Takeaways
- A Wharton-led survey conducted with Gallup and the Walton Family Foundation, of nearly 2,500 United States adults aged 18 to 28 in October 2025, found 79 percent believed AI makes people lazier and 62 percent had concerns that it makes people less smart.
- The same generation leads the workplace in AI adoption. Pew Research Center data published in June found 66 percent of people aged 18 to 29 have ever used an AI chatbot, up from 55 percent in 2024.
- The contradiction is the finding, not a flaw in the survey. Gen Z judges colleagues for using AI while using it heavily themselves, and the researchers describe this as deep ambivalence rather than hypocrisy.
- The researchers put forward two psychological explanations: a bias toward immediate gratification, which makes the short-term benefit of AI outweigh the long-term risk to thinking skills, and the better-than-average effect, which lets people believe AI makes other people lazy but not themselves.
- The report's authors argue employers should not ban AI but embrace the ambivalence, and that workers should outsource the friction in a task, not the thinking that makes the task worth doing.
- The survey does not prove that AI use damages critical thinking in any individual worker. It measures attitudes, and the evidence about actual effects on thinking comes from separate, smaller studies that are suggestive rather than conclusive.
What the Headline Actually Says
Gen Z judges colleagues who use AI, but may paradoxically see it as the key to their own promotion. That is the headline Fortune put on the story, and it is worth reading carefully, because the headline contains two claims, and only one of them is directly supported by the numbers in the article.
The first claim is well supported. Young workers do judge colleagues who use AI. The Wharton-led survey found that 79 percent of respondents believed AI makes people lazier, and 62 percent said they had concerns that it makes people less smart. Those are attitudes about other people's AI use as much as about their own, and the language of the reporting is explicit: Gen Z believes using AI is making their colleagues dumb and lazy. The judgement is not a fringe view. It is the majority position among the young adults surveyed.
The second claim, about promotion, is interpretive rather than numerical. The article does not publish a percentage of respondents who said AI is the key to their promotion. What it describes is a mechanism. As Gen Z grapples with finding or keeping jobs and climbing career ladders, job performance bolstered by an AI boost may hold more appeal than the less tangible threat of losing critical thinking skills. In other words, the promotion argument is an inference from the tension between what young workers believe and what they do: they worry AI makes people lazier, they use it anyway, and the most plausible reason is that it helps them perform. That inference is reasonable, and the researchers endorse it, but readers should know it is not a survey result in the same sense as the 79 percent figure. We flag this distinction in the verification section below, because the headline is stronger than the data behind its second half.
The story itself has an unusual history that is worth stating. A version of it was published on Fortune.com on 30 January 2026, and the version circulating on 15 November 2026 is a republication of that earlier story. The survey at its centre was published earlier in 2026, based on fieldwork completed in October 2025. The findings are therefore about attitudes measured more than a year ago, and the fact that the story is being read again in November says something about how durable the anxiety is, not about a new survey. Nothing in this article treats the findings as fresh fieldwork from November 2026, because they are not.
The Survey: Who Was Asked, and What They Said
The survey was led by researchers at Wharton, the business school of the University of Pennsylvania, and conducted in partnership with Gallup and the Walton Family Foundation. It surveyed nearly 2,500 United States adults between the ages of 18 and 28, and fieldwork was completed in October 2025. The sample is young adults, not only workers, and not only graduates, which matters for how far the findings can be generalised. The attitudes described here are those of a generation, not only of people in offices, although the reporting focuses on the workplace implications.
The headline numbers are striking. Seventy-nine percent of respondents believed AI makes people lazier. Sixty-two percent said they had concerns that it makes people less smart. Those two figures are the core of the story, and they have been reported consistently across every republication of the Fortune article. They describe beliefs about AI in general, not about a specific tool or a specific workplace policy, and they were measured at a time when AI use among young people was already widespread and rising.
The researchers' own summary of the finding is a single word: ambivalence. Benjamin Lira Luttges, a postdoctoral scholar at Wharton who led research for the report, told Fortune: "What we find is deep ambivalence on how Gen Z is thinking about using AI." The ambivalence is not between using AI and refusing it. It is between believing AI is harmful to thinking and using it anyway, heavily, and increasingly. That combination is what makes the finding interesting. A generation that rejected AI would be a simple story. A generation that embraces AI without worry would also be simple. A generation that worries deeply and adopts enthusiastically is a harder story, and it is the one the data supports.
The adoption side of the paradox is documented separately. Pew Research Center data published in June found that 66 percent of people aged 18 to 29 have ever used an AI chatbot, such as ChatGPT, up from 55 percent in 2024. That is an 11-point rise in about a year, and it makes young adults the most AI-experienced age group in the Pew data. The same reporting notes that nearly one-fifth of the generation is worried about AI displacing them at work, yet they lead the workplace in AI adoption. The fear of displacement and the enthusiasm for adoption sit side by side, in the same people, at the same time.
There is one more number that frames the anxiety. Thirty-seven percent of young adults told Pew they believed the impact of AI on them over the next 20 years would be negative, the most pessimistic attitude of any age group surveyed. That is not a majority, and it should not be reported as one, but it is the highest pessimism of any generation, and it sits alongside the highest adoption. The generation most worried about AI's long-term effect on them is also the generation using it most. That is the paradox in its sharpest form.
Why They Judge Others and Use It Themselves
The researchers offer two psychological explanations for the contradiction, and both are worth understanding, because they change how employers should respond.
The first explanation is a bias toward immediate gratification. Lira Luttges speculates that the biggest mental factor informing Gen Z's attitudes is a disposition toward preferring smaller, immediate rewards over larger, delayed ones, a tendency he notes is more prominent in younger, developing minds. His own summary is plain: "There's a legitimate tradeoff between benefits and costs that you get from using AI. Our brains are wired to prefer smaller, immediate rewards versus long-term, delayed rewards." The benefit of AI is immediate and concrete: a task finished faster, a deadline met, a manager impressed. The cost is distant and abstract: a gradual erosion of critical thinking skills that may not show up for years. Faced with that trade-off, most people choose the immediate reward, and young people, whose working lives are dominated by short-term performance pressure, choose it most readily.
That logic also explains why workers use AI even when their employer tells them not to. The reporting notes that young workers use AI even when explicitly told not to, and Lira Luttges offers a direct explanation: even if an employer does not want an employee using AI for certain tasks, those workers, particularly if young, might consider getting their tasks done efficiently as more important, particularly if the risk of getting caught is low. Efficiency is visible and rewarded. Compliance with an AI ban is invisible and unrewarded. When the two conflict, efficiency usually wins, and the worker's calculation is rational even if the employer's policy is not followed.
The second explanation is the better-than-average effect, a well-documented tendency for most people to believe they are above average at a given task, which is statistically impossible for most people at once. Gen Z survey respondents, for example, may see themselves as AI power-users, Lira Luttges said. The logic runs like this: AI could atrophy critical thinking and make other people lazy, but not me, because I use it skilfully, I check its output, and I know its limits. The judgement of colleagues and the confidence in oneself are two sides of the same belief. Other people's AI use is lazy because they lean on it. My AI use is smart because I stay in control. Both halves can feel true at the same time, and the survey data is consistent with that reading, even though the survey does not prove it directly.
There is a third factor that the reporting raises without naming as a formal mechanism: career pressure. As Gen Z grapples with finding or keeping jobs and scaling career ladders, performance bolstered by an AI boost may hold more appeal than the less tangible threat of losing critical thinking skills. This is not a cognitive bias. It is a structural condition. Entry-level jobs are the most exposed to AI, and the generation entering the workforce now faces a labour market in which AI skill is increasingly a condition of employment rather than an advantage. Using AI is not only a choice about productivity. It is a choice about employability, and the fear of being left behind is a stronger motivator than the fear of thinking less sharply.
What the Evidence Says About AI and Critical Thinking
The survey measures attitudes, not effects. It tells us what young people believe about AI and thinking, not whether their thinking has actually changed. The evidence about actual effects comes from separate studies, and those studies are smaller, more specific, and more contested than the survey numbers.
The most cited is a study by researchers from Microsoft and Carnegie Mellon University, published last year, which found among more than 300 knowledge workers that leaning too much on AI tools like ChatGPT was associated with diminished critical thinking skills. The study suggested that even using AI for low-stakes tasks such as proofreading can lead to significant negative outcomes in high-stakes contexts, like writing legal documents. That finding is striking, and it is the study the Fortune article leans on most heavily. But its limits should be stated. It involved just over 300 people, it measured association rather than proving causation, and it relied on self-reported behaviour and performance. It is evidence that heavy AI reliance and weaker critical engagement go together. It is not proof that one causes the other, and it is not a measurement of how much thinking ability any individual loses.
MIT-led research reached similar conclusions, according to the Fortune reporting, and the two studies together form the basis for the concern that AI use may erode the skills it replaces. The concern is plausible, and it is consistent with what is known about skill atrophy in other domains: abilities that are not practised tend to weaken. But plausibility is not proof, and the research base is still thin. No large, long-term study has yet tracked workers over years to measure whether AI use causes a lasting decline in critical thinking, and no study has established how much AI use is too much, or which kinds of use are harmful and which are harmless. The honest summary is that there is suggestive evidence of a risk, and no conclusive measurement of its size.
That uncertainty is itself part of the story. The survey respondents are worried about a harm that science has not yet quantified, and they are adopting a technology whose long-term cognitive effects are unknown. That is a reasonable position to hold, and it is also an uncomfortable one, because it offers no clear rule for how much AI use is safe. The researchers' answer is not a limit or a ban. It is a distinction between kinds of effort, which we examine below.
What Employers Should Do, According to the Researchers
The report's authors argue that employers should not ban AI, but rather embrace ambivalence toward it. That recommendation follows directly from the data. A ban would conflict with how the most AI-experienced generation actually works, it would be widely ignored, as the reporting shows it already is, and it would deprive the organisation of the productivity gains that AI offers. Embracing ambivalence means something more specific than tolerating AI use. It means designing work and training around the tension the survey reveals, rather than pretending it does not exist.
The most practical guidance in the article comes from Lira Luttges himself, and it is a distinction between two kinds of effort. "For every task, there are two kinds of efforts," he said. "There is effort that is germane to the task, that is intrinsic to the thing that you're doing, and that kind of effort is the effort that you put in, and gets translated into learning. But there's a lot of effort that is just there, that's just like friction, that doesn't really teach you anything." His conclusion is a sentence that deserves to be quoted in full, because it is the most actionable line in the whole story: "You should outsource the crap, not the craft." The idea is that AI should take over the friction, the formatting, the searching, the first draft, the routine, while the worker keeps the thinking that makes the work worth doing and that builds the skill to do it again.
That distinction has immediate implications for how managers assign work. If a junior employee uses AI to draft a report and then reviews, corrects and improves it, the germane effort is preserved, because the review is where the learning happens. If the employee pastes the AI output without reading it, the germane effort is gone, and the skill that the task was supposed to build is not built. The manager's job is not to police which tool was used. It is to design the review step so that the thinking still happens, and to check the output in a way that makes silent copying visible. That is harder than a ban, but it is the only approach that survives contact with how people actually work.
There is also a longer-term warning in the article that managers should take seriously. Some future-of-work experts, including Mark Beasley, professor and director at North Carolina State University's Poole College of Management, believe a critical thinking gap, not an AI skills gap, will pose a serious threat to organisational pipelines and business operations. Beasley told Fortune that the threat AI poses to entry-level jobs could mean insufficient training and experience for middle and eventually upper-tier positions in the near future. His summary is blunt: "The biggest risk organizations face is just being stagnant." The logic is that if entry-level workers never do the hard thinking, because AI does it for them, they will not develop into the managers and specialists the organisation needs later. That is a pipeline risk, not an individual one, and it is the strongest argument in the article for deliberate training rather than passive adoption.
One finding offers cautious reassurance. According to the survey, respondents who reported using AI more frequently worried less about its impact on intelligence and motivation, indicating that AI anxiety may resolve over time as familiarity grows. Lira Luttges suggested that as long as workplaces are intentional about how they implement AI, the technology will not have a significant impact on critical thinking. That is an optimistic reading, and it depends on the word intentional doing a lot of work. Familiarity reduces anxiety, but it does not by itself preserve skill. The reassurance holds only if the familiarity comes with deliberate practice of the thinking that matters, which returns us to the same conclusion: the outcome depends on how work is designed, not on how much AI is used.
What This Means for Young Workers Themselves
The survey is about Gen Z, but its most useful lessons are for the young workers it describes, and those lessons are practical rather than moral. The first is that the judgement is mutual. If you judge colleagues for using AI, they are judging you for the same thing, and the 79 percent figure means the judgement is the norm, not the exception. That has consequences for how openly you use AI at work, and for whether you disclose it. In a workplace where most people believe AI makes people lazier, visible AI use carries a reputational cost, even when the work it produces is good. Disclosing AI use, showing your working, and demonstrating that you checked and improved the output are ways to pay that cost down. Silent use invites the assumption that the machine did the thinking.
The second lesson is about skill maintenance. The Microsoft and Carnegie Mellon finding, and the MIT-led research behind it, both point in the same direction: the risk is not using AI, it is leaning on it. The workers who will be most valuable in five years are not the ones who avoided AI, and not the ones who used it for everything, but the ones who used it for friction and kept the craft for themselves. That means doing some work without AI deliberately, checking AI output against your own judgement rather than accepting it, and being able to explain and defend what you produce. Those habits are the difference between using AI as a tool and being replaced by it, and they are habits that can be practised now.
The third lesson is about the promotion question that the headline raises. If AI use does help performance, and performance drives promotion, then the workers who use AI well will advance faster, and the workers who refuse it on principle may fall behind. The survey does not prove that AI use causes promotion, and no reader should treat it as career advice in that form.
For readers thinking about the wider question of which jobs AI can and cannot do, our analysis of what jobs AI cannot replace examines the roles where human judgement, physical presence and trust still matter, and where they are likely to keep mattering. The Gen Z survey is, in part, a story about that boundary: young workers sense that something valuable is at stake when AI does the thinking, and they are trying, without a clear map, to decide where the boundary lies. Our beginner's guide to AI workflow automation is a practical starting point for anyone who wants to use AI at work deliberately, with the friction removed and the thinking preserved, rather than drifting into heavy reliance by default.
What We Could Not Verify
- The promotion claim in the headline. The Fortune article does not publish a percentage of respondents who said AI is the key to their own promotion, and we could not find such a figure in any republication of the story. The promotion argument is presented here as the researchers' interpretation of the tension between belief and behaviour, not as a measured survey result.
- The exact sample size and methodology beyond what is reported. The survey is described as nearly 2,500 United States adults aged 18 to 28, completed in October 2025, conducted by Wharton with Gallup and the Walton Family Foundation. We could not verify the precise number of respondents, the sampling method, the margin of error, or the exact wording of the questions, because the underlying report was not available to us at the time of writing.
- The precise figures for workplace adoption by generation. The reporting states that Gen Z leads the workplace in AI adoption, citing CNBC, but we could not verify the underlying adoption rates by age group, so that claim is reported as stated in the coverage rather than as a figure we independently confirmed.
- The causal claim that AI use diminishes critical thinking. The Microsoft and Carnegie Mellon study of just over 300 knowledge workers found an association, and MIT-led research reached similar conclusions, but neither study proves causation, and no long-term study has yet measured the size of any lasting effect.
- Whether the 79 percent and 62 percent figures have changed since October 2025. The fieldwork is more than a year old, and attitudes may have shifted as AI use has become more common. No newer wave of the same survey was available to us.
- The full text of the original Fortune article. The version we verified is a republication of a story first published on Fortune.com on 30 January 2026, and quotations are drawn from that republication. We did not read the original Fortune page directly.
Sources
- Fortune, via The Fifth Skill, Gen Z judges colleagues who use AI, but paradoxically may see it as the key to their own promotion, 15 November 2026, https://thefifthskill.com/gen-z-judges-colleagues-who-use-ai-but-paradoxically-may-see-it-as-the-key-to-their-own-promotion/
- Fortune, via Europe Says, Gen Z judges colleagues who use AI, but paradoxically may see it as the key to their own promotion, 15 November 2026, https://www.europesays.com/3377259/
- Fast Company, Want to drive AI innovation? Talk about feelings, 12 November 2026, https://www.fastcompany.com/91638717/ai-innovation-work-talk-about-feelings
- What Jobs Can AI Not Replace?
- AI Workflow Automation for Beginners