Gallup's latest survey of Americans aged 14–29 reveals a striking paradox: while half use AI tools like ChatGPT or Microsoft Copilot every week, 42% now believe those same tools harm their ability to think carefully. Excitement about AI dropped 14 percentage points in a single year, to just 22%. Yet weekly or daily usage held steady at 51%. These data, collected from February 24 to March 4, 2026, from 1,572 respondents, paint a picture of a generation that is both deeply reliant on AI and deeply worried it is eroding fundamental skills.

The findings come from Gallup's "Voices of Gen Z" study, one of the most comprehensive looks at how young people in the U.S. view artificial intelligence. For Windows users, IT managers, and anyone navigating the rapid rollout of tools like Microsoft 365 Copilot, the numbers demand attention—not because they signal a rejection of AI, but because they expose a trust gap that threatens both productivity and long-term talent development. This article breaks down what the survey found, why it matters for your daily work, and what concrete steps you can take to use AI responsibly without sacrificing critical thinking.

The Data in Detail: Usage Up, Trust Down

Gallup’s survey uncovers a generation deeply conflicted about artificial intelligence. The headline figure—42% who say AI will harm their ability to think carefully about information—sits alongside similarly stark numbers: 38% believe it hurts their capacity to generate original ideas, and eight in 10 think it is at least somewhat likely that tools designed to complete tasks faster will make learning harder later on. That’s not a fringe view. It’s the majority of young people looking at the very technology they use every day and wondering if it’s working against them.

Among employed Gen Z respondents, the distrust intensifies. Forty-eight percent now say AI’s workplace risks outweigh its benefits, versus just 15% who see benefits as dominant. That 69% majority trusts work done entirely without AI more than AI-assisted work, compared to 28% who favor AI-assisted outputs. These aren’t hypothetical worries; they’re reflected in behavior. A Wharton-led study with Gallup and the Walton Family Foundation found that 74% of Gen Z had used a chatbot in the previous month, yet 16% admitted to using AI at work despite being expressly forbidden to do so.

Numbers from GoTo and Workplace Intelligence’s 2026 Pulse of Work survey, which polled 2,500 global employees and IT leaders, add another layer. Sixty-two percent of Gen Z workers told researchers they treat AI as a crutch, and 40% said they could not function without it. A separate workplace investigation, highlighted in the original CXO Today analysis, found that 44% of Gen Z employees actively undermined their company’s AI rollout—feeding sensitive data into unapproved tools, intentionally slowing adoption, or quietly weakening their own AI-assisted outputs. That isn’t defiance; it’s what happens when people don’t trust that the organization has their long-term interests in mind.

What It Means for You: A Practical Breakdown

For Employees and Everyday Windows Users

If you’re a Gen Z professional using Windows PCs with Copilot or web-based AI assistants, this research likely reflects tensions you feel daily. You know that AI can speed up email drafting, meeting summaries, or code debugging. But you also sense that bypassing the slow, messy process of doing those things yourself might leave you less capable in the long run. Gallup’s finding that eight in 10 think over-reliance will make learning harder validates that instinct. It’s not just you—most of your peers think you’re all quietly losing your edge.

The data also explains the secretive double-play many young workers report: they use AI despite bans or don’t disclose it, fearing they’ll fall behind otherwise. But that secrecy backfires. It creates security vulnerabilities, as the 44% figure on undermining rollouts shows, and it can erode trust with managers. You don’t have to choose between using AI and protecting your skills, but you do need to be intentional. If you’re using Copilot in Word or Excel, try writing a first draft or analysis manually before prompting, to preserve your own thought process. When debugging code with GitHub Copilot, spend time understanding suggestions rather than blindly accepting them. And if your employer lacks an AI policy—56% of IT leaders told GoTo they have none—ask for guidance. Most organizations are still figuring this out, and your questions can help shape a healthier culture.

For IT Managers and HR Leaders

The survey should be a wake-up call for anyone managing Windows-based environments where AI is being adopted. The central insight is this: your youngest workers are both your most prolific AI users and the most skeptical about its long-term value. That skepticism isn’t a barrier; it’s a resource. It means they’re already thinking critically about AI’s impact—something you want to encourage, not suppress.

A knee-jerk ban on tools like ChatGPT or a blanket restriction of Copilot features will likely backfire. As the numbers show, a significant minority of Gen Z workers already use AI against explicit rules. Instead, IT teams need to channel that critical awareness into better governance. Practical steps include establishing clear, written guidelines for approved tools, data-sharing rules, and review processes; training employees not just on how to use AI, but on how to verify outputs, cite sources, and recognize tasks that demand independent analysis; preserving unassisted practice during onboarding, incident response, and code reviews; and measuring quality and analytical reasoning, not just speed or prompt volume.

The GoTo finding that 84% of employees say their company isn’t doing enough to promote responsible use is a clear mandate. For Windows shops rolling out Microsoft 365 Copilot, this means using built-in compliance features like sensitivity labels, data loss prevention, and the Copilot activity dashboard. But technology alone won’t cut it. Acknowledge the tension, and make it safe to discuss.

For Developers and Power Users

Those who use GitHub Copilot or similar tools face a unique risk: automation of core coding tasks can erode the muscle memory that builds deep problem-solving skills. The remedy isn’t to abandon AI, but to deliberately mix assisted and unassisted work. Set rules for yourself—accept only every third Copilot suggestion, or force yourself to understand why a suggestion works before implementing. Schedule periodic “no-AI” sprints where code must be written manually. These habits mirror the concept of “no-meeting Wednesdays” but for cognitive hygiene, and they’re essential for long-term career growth.

How We Got Here: A Technology Rush Outpacing Skill Development

The current angst didn’t emerge in a vacuum. Generative AI tools became widely accessible in late 2022, and Microsoft integrated them rapidly into Windows and Office. The promise was always productivity: Copilot drafts documents, suggests Excel formulas, summarizes Teams conversations. GitHub Copilot promised faster code shipping. The rollout was swift, with little discussion about cognitive tradeoffs.

This isn’t the first deskilling scare. Calculators didn’t eliminate the need to learn arithmetic, but they changed curricula. Spell check didn’t ruin proofreading, but it altered expectations. What’s different now is the breadth of cognitive tasks AI performs—from synthesis and analysis to creative ideation—often without leaving a trace for the user to learn from. The pandemic accelerated digital transformation, pushing millions into remote work and reliance on cloud tools. Gen Z, entering the workforce during or shortly after, encountered AI as standard rather than novel. Their high usage rates reflect not just curiosity but a survival instinct in a job market that increasingly rewards speed.

What to Do Now: Concrete Steps for Windows Users and IT Organizations

Based on the survey insights and expert recommendations, here are actionable steps tailored to different Windows-centric scenarios:

For Individual Users (Windows Home or Pro, using Copilot or web AI):
- Self-audit your AI use. For one day, track every time you use AI. Note whether it replaced a task you could do or expanded your capabilities.
- Practice “AI-free” intervals. Dedicate a work block—say, the first hour of your day—to writing, coding, or problem-solving without any AI assistance.
- Verify and personalize. When Copilot drafts an email or document, re-read it critically. Edit heavily. Add your own insights.
- Turn off Copilot suggestions when learning. In Microsoft apps, you can temporarily deactivate Copilot via File > Options > Copilot settings. This forces you to figure things out yourself.

For IT Departments and Admins:
- Draft and communicate an AI acceptable use policy. Even a one-pager clarifying approved tools (e.g., Copilot, Bing Chat Enterprise) vs. banned tools (public ChatGPT for work data) reduces shadow AI.
- Use Windows and Microsoft 365 security controls. Enable auditing for sensitive file access, set up DLP policies that detect AI-related data exfiltration, and use Microsoft Purview to monitor Copilot interactions.
- Integrate AI training into onboarding. New hire training should include a module on “Using AI without losing your skills,” with concrete exercises.
- Reward deep work. Find ways to recognize employees who produce thoughtful, original analysis, not just fast output.

For Developers and Power Users:
- Pair Copilot with deliberate practice. Set rules: maybe accept only every third suggestion, or force yourself to understand why a suggestion works before implementing.
- Conduct regular “no-AI” sprints. In development teams, schedule periodic sprints where code must be written unassisted to keep skills sharp.

Outlook: Where This Goes Next

The conversation around AI and skill degradation is only beginning. As Microsoft and other vendors push Copilot deeper into Windows—embedding it in Notepad, Paint, and system-level tasks—the line between assistance and dependency will blur further. The next 18 months will likely see a push from regulators and educational institutions for “AI literacy” standards, similar to past digital literacy initiatives. For workplaces, the companies that thrive will be those that turn this generational anxiety into a productive dialogue. Gen Z workers are not Luddites; they’re early adopters with a clear-eyed view of the risks. Organizations that provide structure—clear policies, training that emphasizes judgment, and a culture that values learning over speed—will harness AI’s benefits without hollowing out their talent pipelines. Ultimately, the 42% who fear for their critical thinking may be the very people who save it, if employers listen closely.