High Adoption, Mixed Emotion: What the Research Says About Gen Z, Work, and AI

The dominant theme in recent Gen Z work research is pragmatic idealism. Students and young workers want meaning, but they also want realism: a stable job, decent pay, room to grow, and a life outside work. Deloitte’s 2025 global survey of 23,482 respondents across 44 countries found that Gen Z is focused less on climbing the ladder and more on opportunities to thrive on and off the job — only 6% said their primary career goal is reaching a senior leadership position. Learning and development ranked among the top reasons they chose an employer, and 70% said they develop career skills at least weekly.

Among current college seniors, practical concerns are especially visible. Handshake’s Class of 2025 survey found that desired location mattered more than high starting salary in job-application decisions, and job stability, employer reputation, and schedule flexibility all ranked above hybrid or fully remote work. The same report found that students most often described the labor market as “competitive,” and 56% were somewhat or very pessimistic about starting their careers.

Remote and hybrid preferences are more nuanced than the common claim that young people simply want to work from anywhere. NACE found that the entry-level job market in 2025 was predominantly hybrid or in-person, not fully remote, and that this broadly matched student preferences for some in-person connection. Gallup similarly found in 2025 that Gen Z is the least likely remote-capable generation to prefer fully remote work, at 23% versus roughly 35% for each older generation. Early-career Gen Z professionals appear to move even more strongly toward hybrid once they enter the workplace: NACE found that more than 60% prefer hybrid, but only 38% actually have it.

Gen Z students are not rejecting work. They are rejecting weak trade-offs.

Mental health is not a side issue in this literature; it is a core work variable. Deloitte found that 40% of Gen Z respondents felt stressed or anxious all or most of the time in 2025, with financial worries and workplace issues both contributing. Handshake’s Class of 2024 report found that 80% had experienced burnout symptoms during college, and majorities said work-life balance resources and mental health supports were very important in a future employer. A 2024 Student Voice survey reported by Inside Higher Ed found one in five college students extremely stressed thinking about or preparing for life after college.

Financial pressure is a major driver of work preferences. Deloitte reported worsening feelings of financial insecurity among Gen Z in 2025, while Handshake found that more than half of Class of 2024 seniors expected to carry student loan debt and that nearly 70% of indebted students expected debt to influence which jobs they considered. These pressures also spill into gig work: more than 40% of those seniors expected to pursue freelance, part-time, or gig work after graduation, and nearly a third expected to do it alongside a full-time job.

Diversity and inclusion still matter materially in employer choice. In Handshake’s Class of 2024 research, 77% said they were more likely to accept a job at a company with a diverse leadership team. Deloitte’s 2025 survey showed that Gen Z workers expect managers to foster a positive and inclusive work culture more often than they think managers actually do. For many students, this is no longer an employer-brand issue; it is part of how they judge credibility, safety, and future opportunity.

Career development remains central. Deloitte found extremely high demand for mentorship and practical learning, with 86% of Gen Z respondents emphasizing mentorship and 88% emphasizing on-the-job learning and practical experience. NACE’s early-career survey shows why: Gen Z professionals who had engaged in experiential learning reported faster-than-expected career progression, higher career satisfaction, stronger workplace mentoring and networking, and an average salary advantage of about $15,000 over peers without experiential learning. NACE is careful to note this is an association in a non-probability sample, not a demonstrated cause — but the size of the gap is hard to ignore.

The cleanest summary is this: students are already using AI, and that does not mean they feel secure about what AI will do to their careers. Handshake reported that more than 80% of Class of 2026 students had used generative AI tools, and 57% used them weekly. Gallup found that 57% of U.S. college students were using AI in coursework at least weekly in late 2025, including about one in five who used it daily. HEPI’s 2025 UK survey found AI use among full-time undergraduates had risen sharply, with use increasing across every measured category.

Students often describe AI as a collaborator rather than a substitute for thinking. Handshake’s Class of 2026 report says students primarily use AI as a brainstorming partner or study aid rather than as a content-generation machine. Gallup found the most important reasons students cite for using AI are understanding complex material and saving time, followed by better grades. This complicates the simplistic view that student AI use is mostly about cheating or shortcut-taking.

But optimism is only half the picture. Walton Family Foundation and Gallup reported in 2025 that 79% of Gen Z had used AI tools, yet 41% said AI made them feel anxious and 36% excited; 27% felt hopeful and 22% angry. The same study found that adult Gen Zers were much more likely than school-aged Gen Zers to say AI made them feel anxious — 53% versus 21%. The closer you get to the labor market, the heavier it sits.

Deloitte’s global data reinforce the ambivalence. In 2025, 61% of Gen Z and millennial respondents worried that generative AI would make it harder for younger generations to enter the workforce by automating tasks often assigned to entry-level workers. Yet in Deloitte’s 2026 survey, nearly three-quarters of Gen Z and millennial respondents — 74% combined — reported using AI in their day-to-day work, and largely described it as an accelerant rather than a threat. The tension is important: the same person can believe AI helps them today and worry that it narrows pathways into work tomorrow.

Students’ AI fears cluster around four themes.

2.1 Automation anxiety

Gallup’s U.S. college-student research found that 42% of bachelor’s students and 56% of associate-degree students had given at least a fair amount of thought to changing their major because of AI, and 16% of students said they had already changed their major or field because of AI’s potential impact. BestColleges similarly found that 53% of students were worried about AI’s impacts on the workforce and 27% said AI in the workforce was making them doubt or reconsider their major.

2.2 Ethical discomfort and policy uncertainty

NACE found that 67.1% of Class of 2025 graduating seniors did not use AI in their job search; among nonusers, the leading reasons were ethics concerns, lack of expertise, and fear that employers would know they used AI. Inside Higher Ed’s 2025 Student Voice coverage found that students wanted action on academic integrity, but preferred education on ethical AI use and clearer standardized policies over surveillance-heavy responses like AI detectors. HEPI’s 2025 UK survey likewise found growing use alongside persistent uncertainty about what acceptable use actually looks like.

2.3 Skill-gap anxiety

Walton-Gallup found that 44% of Gen Z students believed they would need to know how to use AI in their future careers and 52% said schools should be required to teach it, yet only 28% said their school explicitly allowed AI use and nearly half said their school had no policy or they were unsure. Gallup’s Heartland study similarly found that only four in ten school-aged Gen Z respondents felt prepared to use AI in future jobs. Pearson and AWS’s 2026 AI readiness research goes further: 53% of employers say finding graduates with the right AI skills is their top hiring challenge.

2.4 Thinking quality and creativity

Walton-Gallup found that nearly half of Gen Z respondents were concerned AI would harm their ability to think carefully about information. Handshake quotes from students echo the fear that overuse becomes a crutch and erodes originality or voice. Academic research points the same direction from an unexpected angle: a 2025 peer-reviewed structural-equation study of 455 Chinese college students found that positive AI attitudes and higher AI literacy reduced job-seeking anxiety — implying the anxiety is not only about the technology but about students’ confidence in understanding and using it.

At the same time, labor-market evidence makes AI literacy hard to ignore. PwC’s 2025 Global AI Jobs Barometer found that workers with AI skills commanded a 56% global wage premium, up from 25% the year before. NACE reported in April 2026 that more than one-third of entry-level jobs now require AI skills, with the share nearly tripling since fall 2025. Employers in that NACE survey were far more likely to discuss AI as a tool for task change and productivity than for outright headcount replacement — just 11% were discussing replacing positions — but the direction is unmistakable: AI capability is moving into baseline employability.

One of the biggest mistakes in talking about Gen Z is treating it as a single psychological type. The recent evidence shows meaningful differences by student status, career stage, field of study, socioeconomic background, degree type, and gender.

Students and early-career workers do not see work the same way. NACE’s comparison of graduating seniors with Gen Z professionals one to three years out found that students placed especially high value on mental-health support, tuition reimbursement, and family-friendly benefits, while those already in the workforce were more likely to want work-from-home ability and hybrid arrangements. Initial student preference for in-person work shifts quickly toward hybrid once they experience work directly.

Field of study shapes AI exposure and emotion — but not in the expected direction. Gallup found that students in business, technology, and engineering use AI most frequently. Handshake found tech majors somewhat less likely than non-tech majors to be worried about AI overall in the Class of 2024, but in the Class of 2025 the share who were highly worried was highest in computer science, followed by humanities. That mix suggests two mechanisms operating at once: higher exposure creates both fluency and sharper awareness of disruption risk.

Gender differences recur. Inside Higher Ed and College Pulse found that men were more likely than women or nonbinary students to say AI was shaping their career plans. Gallup found male college students used AI more frequently than female students, especially daily. Handshake’s Class of 2024 research found women more likely to view salary negotiation as risky. These are not trivial differences; AI-readiness initiatives that assume the same confidence baseline for all students may widen rather than close gaps.

Socioeconomic background matters, and there is a genuine AI divide. HEPI’s 2025 UK survey found that students from higher socioeconomic groups were more likely to use AI for cognitively strategic tasks — summarizing articles, structuring thoughts, editing AI-generated text — whereas students from lower socioeconomic groups were more likely to say they had used AI for none of those. Gallup’s Heartland work similarly found that lower-income and rural students were least likely to say their schools allowed AI use, and especially likely to say schools had no clear rules.

Degree level matters too. In Gallup’s 2026 State of Higher Education work, associate-degree students were more likely than bachelor’s students to rethink their major because of AI, and slightly more likely to have already changed it. AI disruption may be felt most strongly among students closest to direct occupational pipelines, or who perceive less room for experimentation.

Geography shows up in institutional context more than in attitudes. Much of the student-focused evidence comes from the U.S. and U.K. The most globally scaled dataset here — the Digital Education Council’s 2026 survey of 45,398 respondents across 35 countries — found that only 28% of students thought most or many of their assessments matched the skills and judgment needed in an AI-enabled workplace, and 60% worried peers might misuse AI for unfair advantage, rising to 73% in the U.S. and Canada. That points to a broad international problem of trust and curriculum alignment, with North American anxiety running somewhat hotter than the global average.

Where the evidence genuinely conflicts

The first conflict is between student fear and employer optimism. Handshake reported that only 24% of rising seniors thought generative AI would create jobs, while 55% of hiring managers believed it would. NACE’s 2026 employer-side evidence points the same way: AI is reshaping entry-level work far more than replacing it.

The second is between high AI usage and low confidence in readiness. Students use AI heavily, but Pearson’s 2026 readiness work, Walton-Gallup’s 2025 findings, and Gallup’s college-student and Heartland studies all point to the same gap: exposure is not the same as professional fluency. A student can know how to prompt a chatbot and still feel unprepared to use AI responsibly in a workplace with policy, privacy, quality, and reputational stakes.

Exposure is not fluency. Knowing how to prompt a chatbot is not the same as knowing how to be accountable for its output.

The strongest conclusion from this research is that you should think in terms of task resilience, not AI-proof majors. World Economic Forum employer data show both job creation and displacement ahead, with 39% of job skills expected to change by 2030 and AI, big data, and cybersecurity among the fastest-growing technical skills. NACE’s 2026 data show AI skill requirements moving quickly into entry-level hiring. Your competitive edge will come less from choosing a supposedly safe title and more from combining domain strength, technical adaptability, and human judgment.

The practical skill stack is narrower than many students think, but deeper. You do not need to become an AI engineer unless your field requires it. You do need basic AI literacy: knowing what leading tools can and cannot do; writing better prompts; checking outputs for error, bias, fabrication, and style problems; understanding privacy and IP boundaries; and showing you can insert AI into a workflow without letting it take over your thinking. Deloitte found that Gen Z workers see AI skills as important — but communication, leadership, empathy, networking, time management, and industry-specific knowledge still rank higher.

Resumes and portfolios: show evidence, not tools

A weak AI-ready resume lists “ChatGPT” or “AI tools” in a skills section. A strong one shows evidence. Handshake reports that resumes mentioning AI have doubled since 2022, so generic tool familiarity is no longer a differentiator — the ability to use AI well in context is. Your portfolio should show projects where you used AI to speed research, improve analysis, test ideas, draft alternatives, or support decision-making, and where you added human judgment, verification, and revision.

A good portfolio entry now has four parts: the problem, the tool or workflow used, the human decisions you made, and the measurable result.

That framing signals exactly what employers are asking for: AI as a complement to judgment, not a substitute for it.

Interviews: three things to be ready to say

NACE’s findings on job-search AI use show that many students are uneasy precisely because they do not want to appear dishonest or overreliant. Turning that anxiety into a clear account of your standards is an advantage. Be ready to speak to three things:

1. How you use AI productively. Where in the workflow it earns its place — brainstorming, comparison, first-pass drafting, synthesis.
2. How you verify and edit its outputs. What you check, what you rewrite, how you source claims.
3. How you handle ethics and boundaries. Company policy, sensitive data, when you escalate to a human reviewer.

A strong answer sounds like this in substance: “I use AI early in the workflow for brainstorming, comparison, or first-pass drafting, but I do not treat its output as finished work. I verify claims, rewrite for audience and voice, and document sources. In roles involving sensitive data or customer-facing decisions, I would follow company policy and bring escalations to a human reviewer.” That directly addresses employer concerns about quality control, ethics, and judgment.

Above all: get real experience

The NACE early-career data are some of the strongest evidence here. Internships, co-ops, research, client projects, and microinternships are associated with higher career satisfaction, stronger mentoring networks, and materially higher early-career salaries. In an AI-shaped market, experience matters not only because it adds lines to a resume, but because it gives you concrete examples of judgment under uncertainty — exactly the thing AI cannot prove on your behalf.

The most important habit is iterative learning, not one-off certification. WEF expects large skill shifts by 2030, and Deloitte, Gallup, Handshake, and Pearson all point to the same reality. Treat AI fluency the way previous cohorts treated spreadsheet fluency or internet research — necessary, normal, but only valuable when attached to sharp thinking and real output.

Fifteen studies carry most of the weight in this piece. They differ substantially in method, sample, and geography, and those differences matter when reading any single number.

The research is good enough to show direction, but the limitations are real. Most student-facing datasets are from the U.S. and U.K., so claims about “Gen Z globally” should be made carefully. Much of the evidence is self-reported survey data rather than behavioral or longitudinal evidence. Platform-based surveys such as Handshake’s are highly informative about active job seekers but may not represent all students. And although many surveys capture anxiety and optimism, far fewer track whether those attitudes actually predict later career outcomes.

Sentiment also moves fast. The Walton-Gallup emotion splits reflect March 2025 fieldwork, and later waves have shown more negative sentiment. Any snapshot of how students feel about AI has a short shelf life; the structural findings — about experiential learning, complements to AI skill, and the gap between exposure and fluency — hold up better across waves.

There is also a genuine disagreement about whether AI is more likely to compress early-career pathways or upgrade them. Student surveys emphasize threat; employer surveys emphasize task redesign and opportunity; labor-market analyses show rising demand and wage premiums for AI-related capability. The best current interpretation is not that one side is wrong, but that transition costs are real and uneven.

The upside goes first to students who have guidance, practice, and the ability to show disciplined use.

That last clause is the whole argument compressed. Not the students who used AI the most. The ones who can show what they did with it, and why they trusted the parts they trusted.