AI Reliance Study

A Student-Led Guide to Using AI Wisely in School

We talked to students, ran experiments, and built a simple framework to help you use AI like ChatGPT as a learning tool, not a crutch. No complicated rules just practical advice.

Our Mission

To help students use AI without losing their own skills.

Goal

Create a simple, practical guide for responsible AI use, based on what real students told us.

Research

We interviewed 8 students and ran an experiment with 8 more to see how a framework changes behavior.

Framework

5 evidence‑based principles to keep you learning and thinking, even when you use AI.

Impact

Our framework helped students produce better work and be more thoughtful about AI.

8
Students Interviewed
8
Experiment Participants
5
Evidence‑Based Principles
8
Months of Research

Meet the Team

Four students from Carleton University who wanted to make AI easier to understand.

Erin Van Der Pouw Kraan

Lead Researcher

Interviews Ethics

Led the interviews and made sure our study was fair and followed the rules.

Victor Oikawa Lopes

Research Coordinator

Framework Data

Turned interview responses into our 5 guiding principles and planned the experiment.

Mohammad Al-Saadi

Technical Lead

Website Charts

Built this website and made our data easy to see with charts.

Mustafa Ali

Technical Coordinator

Research Analysis

Also helped with the website and worked on the poster for the project.

Our Journey

Sept 2025

Started the Project

Picked our topic and figured out what questions we wanted to answer.

Oct 2025

Got Ethics Approved

Got the green light from Carleton to safely interview students.

Dec 2025

Interviewed Students

Talked to 8 students about their real thoughts on AI.

Jan 2026

Built the Framework

Created 5 evidence‑based rules based on interviews and research.

Feb 2026

Ran the Experiment

Tested our framework with 8 students doing real school tasks.

Apr 2026

Shared Our Findings

Presented everything at the Carleton Capstone Fair.

Our Supervisor

James Brunet

Professor, School of Information Technology

Carleton University

Helped us stay on track, design our study the right way, and turn our ideas into real research.

The AI Learning Framework

A research‑backed approach to using AI wisely, developed from student interviews and validated through experimentation.

How We Built This Framework

Phase 1
Student Interviews

We talked to 8 students about their AI usage, concerns, and experiences. Their responses shaped our initial principles.

Phase 2
Thematic Analysis

We coded interview responses and identified 5 core themes that students cared about most.

Phase 3
Literature Review

We validated our themes against peer‑reviewed research on AI literacy, accuracy, and learning outcomes.

Phase 4
Experimental Validation

Tested the framework with 8 students. Those using it produced higher quality work and showed more thoughtful AI use.

01

Understand the “why” (AI literacy)

Know how the AI generated its response – don’t just copy.

📘 What the research says

Gong et al.: Even students exposed to AI lack understanding of AI ethics.
Boscardin et al. (2024): Accuracy, bias, and ethical use must be considered.
Harvard/Stanford RCT (2025): Using AI as a “scaffolded tutor” doubles learning gains.

🗣️ What students told us

“I use AI to explain concepts in a more human way, but I don’t always know if I can trust it.”

Try this: Instead of “what’s the answer?” ask “walk me through the steps to solve this.”

02

Always fact‑check

AI invents sources and details – especially for niche topics.

📘 What the research says

Johnson et al. (2023): Only 57.8% of AI medical answers were “nearly all correct.”
Linardon et al. (2025): 65% of GPT‑4o citations unreliable; for niche topics fabrication jumps from 6% → 29%.

🗣️ Student insight

“AI gave me a perfect quote with an author – but the paper didn’t exist.”

Try this: Verify every citation on Google Scholar.

03

Don’t form bad habits

Consider what you already know before turning to AI.

📘 Based on interview data

From Q6: overuse and dependence were the main concerns. Students said consulting AI before thinking twice can weaken your own skills.

🗣️ What they said

“Dependence is the biggest negative – I’m worried I’ll lose my ability to write.”

Try this: Work on a problem for 10–15 minutes on your own first.

04

AI is a tool, not a crutch

Be mindful of its limitations – it’s up to you how you use it.

📘 Categorized from Q1 & Q4

Students find AI useful for: repetitive work, rewording, idea generation, summarizing, creating study guides.
Student analogy: “AI is like a lighter. It can light a candle or burn down a building.”

Try this: Use AI for brainstorming, then write the final draft yourself.

05

You own the consequences

The repercussions apply to you, not to the AI.

📘 From Q1 & Q2

Every student knows the academic integrity rules, but only half personally care – the rest just avoid getting caught. “Individuals have found ways to cheat the system – consider if you’re cheating yourself.”

Remember: If you misuse AI, you miss out on learning.

What We Found

Real data from 8 Carleton students – full demographics from our interviews.

Participant Snapshot

We interviewed 8 undergraduate students (Bachelor's level) from Carleton. Here’s who they were:

7
Men
1
Woman
21-23
Age range

Ethnicity: 2 White, 4 Asian, 1 Middle Eastern, 1 Black or African American.
Field of study: 8/8 in Science, Engineering, IT.
CGPA (12‑point scale): ranged from 6.0 to 11.5.

Academic integrity: do you care?

Split: 4 follow rules because they want to learn, 4 just avoid trouble.

Perception of AI

Most see AI as helpful, but dependence worries remain.

AI usage frequency

5 frequent, 2 rare, 1 consistent (daily).

Self‑reported reliance on AI

Only 2 of 8 consider themselves reliant.

The Experiment: Does the Framework Work?

We put our 5 principles to the test with real students doing real tasks.

8 Students

4 got the framework, 4 didn't.

6 Tasks

Summarizing, writing, debugging, scheduling.

About 1 Hour

Each student worked through tasks at their own pace.

So, Did It Help?

Yes – students with the framework were more thoughtful and produced better work.

Better Quality Work

Blind evaluators rated the work from the group with our framework higher.

More Thoughtful

They were more intentional about when and why they used AI.

Slower, But Smarter

Students without the framework finished faster, but quality suffered.

The Bottom Line

Our 5 principles made a difference. Students who had them thought more about their choices. They used AI as a tool to help them learn, not just to get the task done.

Comparison: Framework vs. No Framework

Resources & Downloads

All our materials, free for you to use.

The 5 Rules Poster

A simple, printable poster with our framework. Perfect for your study wall.

Download PDF

Full Research Report

All the details: methods, interview questions, experiment tasks, and data analysis.

Download Report

Capstone Slides

The slides we used to present our project at the fair.

Download Slides

Study Materials

Get in Touch

Questions about our research? Want to collaborate? Reach out!

Email Us Directly

erinvanderpouwkraan@cmail.carleton.ca

victoroikawalopes@cmail.carleton.ca

mohammadalsaao@cmail.carleton.ca

mustafaali6@cmail.carleton.ca

Our Supervisor

James Brunet

JamesBrunet@cunet.carleton.ca