Student Metacognition Guide

Many students use AI in their coursework, and it isn't always clear whether it's helping them learn or getting in the way. This page offers a short guide you can share directly with your students, along with some background on why it matters.

How to use this page: The first section is for instructors. The second section is written for students, so you can share it with them directly. It introduces metacognition briefly and offers four questions to ask themselves whenever they use AI. Students who choose not to engage with AI tools don't need it, but for anyone using AI in their academic work, it can help. To share it, send students this page or download the guide to share as a file.

Why Metacognition Matters Now

Educational research on generative AI's impact on learning is still developing. Here's what stands out so far (see AI and Learning for more):

  • Some studies find that AI use can erode critical thinking (Gerlich, 2025; Bastani et al., 2025; Kulal, 2025), while others find it can improve learning outcomes (Ma & Zhong, 2025; Qu et al., 2025; Wu et al., 2026).
  • Course and assignment design play a critical role in shaping whether AI functions as a learning tool or a disruption to learning (Qu et al., 2025; Xu et al., 2025; Joo, Han, & Park, 2026).
  • Fluent AI output can create an illusion of learning. Because large language models produce polished, confident responses, students may mistake the quality of the output for evidence of their own understanding.
  • AI literacy and metacognitive habits are emerging as two of the most important factors in helping students use AI productively while maintaining their critical thinking skills. The Course & Assignment Design page offers teaching moves that support both.

What is Metacognition?

Metacognition, or “thinking about thinking,” is the awareness of—and ability to direct—one's learning. It involves planning how to approach a task, tracking progress, evaluating outcomes, and drawing on what one knows about the subject, effective learning strategies, and oneself as a learner.

Using AI Without Losing Learning

If you use GenAI tools (such as Gemini, ChatGPT, Copilot, or Claude) for coursework, even occasionally, this guide is for you. It's about something more than whether you're allowed to use AI: it's about whether it's helping you learn or getting in the way.

Starting Point

Using AI to submit work that isn't yours is academic dishonesty, and that hasn't changed. But many students are asking a second question that doesn't have a clean answer yet: Does using AI help you learn, or does it weaken your critical thinking? It depends heavily on how you use these tools. What's clearer is this: some patterns of AI use undermine learning, and others can support it. This guide is about helping you tell the difference while you're using these tools.

What is Cognitive Offloading?

When you use AI, you're letting a tool carry part of the thinking you'd otherwise do yourself, like using GPS to find your way. Researchers call this cognitive offloading.

  • Some offloading is harmless. Using AI to fix formatting or check spelling may not cost you much (unless that's part of what you're trying to learn).
  • The real risk comes when you offload the parts of a task that are the actual point: synthesizing ideas, building an argument, developing your own reasoning. AI's answers can feel like understanding because they're fluent and confident, even when they're wrong.

You can use AI to extend your thinking, not shortcut it. The tool itself isn't what matters most; how you use it is. The lists and questions below can help you tell the difference in the moment.

Not All AI Use is the Same

One kind of use builds your capability. The other substitutes for it. 

These examples can shift depending on your course and the learning objectives your instructor has set, so treat them as a starting point, not a universal rule.

Using AI this way may hurt your learning:

  • Asking AI to write your argument or thesis
  • Copying AI's analysis as your own thinking
  • Generating ideas with AI without evaluating them
  • Submitting AI output with only light editing
  • Asking AI what to think, not how to check your thinking

Using AI this way may support your learning:

  • Using AI to give feedback on your grammar or formatting
  • Asking AI to explain a concept, then verifying it yourself
  • Using AI to brainstorm starting points that you then evaluate
  • Asking AI to challenge your argument by finding its weak spots
  • Using AI to check your reasoning, not replace it

Questions to Ask Yourself Every Time You Use AI

These questions are the thinking that turns AI from a shortcut into a learning tool. Try asking them out loud or writing your answers down until they become a habit.

PAUSE, before you start! AI tools are fast and sound confident, which makes them tempting to reach for before you've decided you actually need them. Before opening a tool, pause and ask: Do I need AI for this, or am I reaching for it out of habit? If you're not sure, that hesitation is useful information. Treat it as a cue to start with what you already know before bringing AI in. If you've already paused and done some initial thinking, good. That's exactly what the first question below builds on.

What do I already know before I ask AI?

  • Why this matters: If you can't answer this, you may have outsourced the thinking. What you already know is what lets you judge AI's response.

  • Try this: Before opening AI, spend a few minutes freewriting, or jot down two or three sentences about what you already know about the topic.

What will I keep, change, or reject from what AI produced, and why?

  • Why this matters: This is the step where learning happens. Deciding what to keep, rewrite, or throw out requires you to think, which is exactly what builds the skill.

  • Try this: Highlight every part of the AI output you're using. For each one, note "keep," "change," or "reject," and write down why.

What is still unclear, and how will I find out?

  • Why this matters: AI produces confident answers even when it's wrong. If something doesn't fully make sense to you, that's a signal to investigate. Don't mistake fluent output for accurate output.

  • Try this: After reviewing AI's output, write down one thing you're still uncertain about. Then look it up in another source.

Could I explain this to someone else without AI's help?

  • Why this matters: This is the real test of whether you've learned something or just borrowed it. If your instructor asked you to explain your reasoning right now, could you do it?

  • Try this: Close the AI tool and explain the key idea of your work out loud in a few sentences.

Remember! AI that does your thinking for you produces work that looks good but leaves you less capable. AI that you direct, question, and verify may build both the work and the skill. Learning isn't a product. It's a process you keep investing in, and the struggle is often the point, not something to shortcut.

References