What Students Are Telling Us About AI—and Why Schools Should Listen

Summary for Educators

Based on Jason Gulya
“3 Student Reflections on Gen-AI and What We Can Learn From Them”
The AI Edventure, July 1, 2025

THE BIG IDEA

Much of the conversation about generative AI in schools is being led by technology companies, researchers, policymakers, and educators.

Jason Gulya argues that one critically important voice is often missing:

students themselves.

He examines reflections from three high school students—William Liang, Olivia Han, and Mary Ruskell—whose experiences with generative AI raise three very different concerns.

One sees AI as the logical shortcut in an educational system dominated by grades. Another worries that repeated reliance on AI is gradually replacing her own writing voice. The third describes something even broader: growing uncertainty about whether anything encountered online can be trusted.

Taken together, their experiences suggest that the challenge facing schools is not simply determining whether students should be allowed to use AI.

Schools must understand how AI is changing students' relationship with learning, authorship, effort, truth, and even trust itself.

KEY TAKEAWAYS

• For some students, assignments have become obstacles rather than learning opportunities. High school student William Liang describes assignments as logistical hurdles that students must get past. If the primary goal is earning a grade, generative AI becomes an extremely efficient shortcut.

• AI may be exposing a deeper problem with grading. Gulya argues that schools have spent decades signaling that grades matter enormously. If students believe the objective is simply obtaining an A, it should not surprise educators when they embrace tools that accomplish the task faster.

• Convenience can slowly become dependence. Sixteen-year-old Olivia Han initially viewed AI as a way to work more efficiently. Over time, however, she realized that AI-generated language was beginning to replace her own voice. She found herself wondering how AI would phrase something before trusting her own words.

• Writing is about more than producing polished prose. Han's reflection reminds educators that writing helps students discover ideas, develop judgment, and construct a personal voice. If AI performs too much of that intellectual work, students may produce better-looking writing while developing less as writers.

• AI raises questions about reality itself. Student Mary Ruskell describes the difficulty of navigating a digital environment in which images, statements, and interactions may be authentic—or artificially generated. The resulting uncertainty can affect not only schoolwork but also students' relationships and their trust in other people.

• Students do not experience AI in the same way. Gulya's three examples reveal very different relationships with the technology. Some students see efficiency. Others see dependency. Still others see an erosion of trust. That variation argues against simplistic school policies that treat every student and every AI use identically.

WHY IT MATTERS

The most important lesson for school leaders may be that AI is not merely a cheating problem.

Consider William Liang's observation.

If students see assignments primarily as hoops they must jump through to earn grades, AI does not create that transactional attitude.

It simply makes the transaction easier.

That should force educators to examine the assignments themselves.

Would students still see value in the work if there were no grade attached?

Does the assignment require genuine thinking, discussion, revision, investigation, or personal judgment?

Or is it largely something students complete because adults require it?

Olivia Han raises a different issue.

Schools understandably want students to use technology efficiently. But efficiency can become problematic when students begin outsourcing the very cognitive struggle that produces learning.

A student struggling to find the right sentence may appear inefficient.

Yet that struggle may be exactly how the student learns to write.

Mary Ruskell presents perhaps the most profound challenge.

For generations, schools have taught students to evaluate sources.

AI raises the stakes dramatically.

Students now live in an environment where photographs can be fabricated, videos manipulated, voices cloned, social interactions automated, and authoritative-sounding text generated instantly.

Media literacy therefore becomes something larger than determining whether a website is credible.

Students increasingly need to ask:

Is this real?
Who created it?
Why was it created?
What evidence supports it?
And what would convince me that it is true?

LEADERSHIP ACTIONS

Ask students how they are actually using AI.
Do not rely exclusively on adult assumptions. Conduct student focus groups, surveys, advisory discussions, and classroom conversations.

Examine transactional assignments.
Ask teachers to identify work that students are likely to see simply as something to “get done.” Redesign those assignments around authentic thinking, discussion, application, and reflection.

Protect productive struggle.
Establish places in the curriculum where students must think, draft, calculate, debate, or solve problems before turning to AI.

Teach students to preserve their own voices.
Have students compare AI-generated writing with their own work and discuss what makes writing sound genuinely personal.

Make AI literacy part of media literacy.
Students should learn how synthetic images, deepfakes, fabricated sources, and AI-generated text can undermine confidence in information.

Listen before regulating.
AI policies written without meaningful student input may fail to address how students actually experience the technology.

LEADER REFLECTION

Schools frequently ask:

“How do we keep students from using AI improperly?”

These students suggest that a better set of questions might be:

Why are students eager to outsource certain assignments?

Which parts of learning should remain difficult?

How do we help students preserve their own voices?

And how do we prepare young people to live in a world where seeing something no longer guarantees that it happened?

Perhaps the most important lesson from Gulya's article is remarkably simple:

Before schools decide what AI should mean for students, we should spend much more time asking students what AI already means to them.

Original Article

------------------------------

Prepared with the assistance of AI software OpenAI. (2026). ChatGPT (5.2) [Large language model]. https://chat.openai.com