We often hear about AI and student cheating. This is a valid concern. But I’d like to highlight ways that we can use AI for deeper mastery. In this post (and podcast), I explore how we can move beyond using AI for quick answers and instead harness it for deeper learning. I share ten practical strategies—ranging from prompt engineering and concept mapping to simulations, Socratic dialogues, and Feynman-style teach-backs—that help students engage more critically and reflectively. Whether students are using AI as a tutor, a thought partner, or a simulator, the goal is the same: to avoid shallow thinking and support true mastery through intentional, creative, and metacognitive use of these powerful tools.

Listen to the Podcast

If you enjoy this blog but you’d like to listen to it on the go, just click on the audio below or subscribe via iTunes/Apple Podcasts (ideal for iOS users) or Spotify.

 

Technology Makes Learning Faster and More Efficient

When I was in the eight grade, I created a slide presentation with recorded audio. Research involved card catalogs, books, and navigating the dark art of microfilm and microfiche. When I interviewed experts, I had to keep a stop watch to see how much money I was spending per minute on long distance phone calls. Making a single slide involved getting adequate lighting, making sure the spacing was perfect, then snapping a picture on a camera. Once the film roll was complete, I would take it to a drug store, have an amazingly cylindrical scoop of Thrifty’s ice cream, and wait a week for my slides to be developed. I recorded my script on giant reel to reel magnetic tape and editing involved a razor blade and Scotch tape.

Now, I draw my slides and color them with Photoshop. I pull from a library of free photos. I record the audio in my office. I interview experts via Zoom and look through journal articles on Consensus (where I get an AI explanation) followed by Google Scholar. Everything about the process has grown faster and cheaper.

As we think about this era of generative AI, it’s easy to see just how fast and efficient the learning process might become in the future. Instant answers. Targeted tutoring. Quality feedback in seconds. I’ve already seen how students can use AI tools to fix audio issues and create jump cuts for videos.

But there’s a potential cost here. If we aren’t careful, we fall into cognitive atrophy, where we allow the algorithms to do the thinking for us. I struggle with spatial reasoning because I have allowed the map app on my phone to do all navigation for me. I worry that the same thing might occur when students use AI as the starting place for their writing (especially given the reality that we often learning through writing rather than write after learning).

If we aren’t careful, students can end up using AI to engage in cursory, shallow approach that leads to cognitive atrophy. However, we can also use these tools in a way that leads to deeper mastery. So, I’d like to highlight some of the ways that we can use AI to deepen the learning process. Note that in some cases, students might be using AI as more of a tutor. In other moments, they might be using AI as a thought partner. Still in other moments, AI functions more like a simulator.

 

Ten Ways to Use AI for Deeper Learning

The following are ten practical strategies you can use for developing deeper mastery with AI.

Strategy #1: Slow Down with Prompt Engineering

Over the last year, I’ve had the opportunity to work with educators who are incorporating generative A.I. into their classroom practice. Whether it’s working with the pre-service teachers at my university or in working with current teachers in the workshops and professional development I lead, I have been inspired by the creative ways teachers are using generative A.I. for student learning. I’m struck by the intentionality and creativity inherent in the process. But I’ve also noticed a surprising trend. Students don’t always know how to craft relevant prompts for the A.I. chatbots. They tend to use chatbots in a rapid-fire way that’s more reminicent of the way they scroll through social media or engage in a video game.

When students move too quickly with a chatbot, they tend to assume the AI tools contain a certain contextual knowledge that they lack. Because of the ELIZA Effect, students often treat the A.I. as more human than it actually is and they forget that the seemingly intelligent A.I. sometimes lacks basic understandings around context clues and empathy.

As they move on, many students merely copy and paste the answers. They don’t engage in deeper analysis. They rarely engage in fine-tuning, where they adjust the prompts through tiny tweaks and experimentation. In some cases, students don’t even read the whole answer. Often, students get into a narrow focus and they don’t ask unrelated questions that might lead to deeper learning. They also tend to get a tunnel vision where they only use one A.I. tool and don’t seek out variations on the answers they receive. In other words, they become rigid in their thinking.

This is why it helps to engage in prompt engineering. I developed the FACTS Cycle as a way to slow students down so that they can be more intentional in what they craft and then analyze what they read. Here is a brief overview of the process:

Here’s how it works:

• Formulate the question: What is it that you want to know?
• Acquire the AI Tool: Look at the pros and cons of various tools to see what works best for the question you have.
• Create Context: Use a RAFT to spell out the role you are taking, the audience for the AI, the format (bullet points, Q&A, spreadsheet, etc.), and the tone (which also includes the temperature of the AI)
• Type the Prompt: Type it into the AI chatbot
• Scrutinize Results: Examine the factual accuracy, the bias, and whether or not it actually answered your question.

This prompt engineering process is a solid start. And honestly, that RAFT element makes a huge difference. But what does it look like to use this process to develop deeper learning?

 

Strategy #2: Using AI as a Deep Conceptual Study Tool

Step 1: Make a quick, cursory read of your work from the last week, unit, or semester. As you go, fill out a t-chart with “what I know” and “what I don’t know.”

Variation: You might even use a standards-based assessment grid to self identify your mastery level of each of the standards.

Step 2: Train a chatbot on your current work. It can help to use a custom chatbot but you can use a basic generative AI tool like Gemini, ChatGPT, Co-Pilot, or Claude. A starting prompt to use is:

I am a grade _______ student. I am learning about _____________. I would like you to review ________ and look for trends in what I know and don’t know. Put it into an easy to understand t-chart.

Variation: The current standards we are learning about are __________. I will be submitting this standards-based assessment grid. Fill out the chart with my mastery level based on the work I have done.

Step 3: Review the AI report of what you know and what you don’t know.

Step 4: Compare and contrast what you see and what the AI tool has analyzed. What are some similarities and differences? What are some things it identified that you missed? What did it miss? Do you agree with the assessment?

Step 5: Select one key concept or learning target that you have not mastered and ask the AI to generate an informational text explaining it. A starting prompt might be:

I would like you to create a (quantify number of paragraphs) informational text about _________. Explain it to me in simple terms. (optional: explain it like I am ______ years old). The tone should be (specific tone – example would be “academic but engaging”). Be sure to keep a low temperature focused on accuracy. Be sure to include front-loaded, easy to understand content vocabulary at the top of the informational text.

Step 6: Take handwritten notes on key ideas from the informational text and jot down a few questions you have.

Step 7: Ask clarifying questions to build your own background knowledge focused on any part of the information you don’t understand.

Step 8: Ask the AI to test you on what you just learned. An optional prompt might be:

Create a series of multiple choice tests that I will answer one at a time. It should progress from basic to move advanced as I get the answers correct. Each time I get the answers incorrect, please explain the correct answer and why I might have gotten it wrong. Once I have answered challenging questions, switch to critical thinking questions that are open-ended and not multiple choice.

Step 9: Handwrite a summary of what you have learned.

Repeat steps 5-9 on additional days with a focus of 30-45 minutes of study time. When you move to step 8, you might also ask it to test you on information from the previous days.

 

Strategy #3: Concept Map Comparison

This is a strategy that’s just now emerging as AI moves toward a place of deeper reasoning. I wouldn’t recommend using it just yet. But the core idea is to use AI as a thought partner with conceptual understanding.

Step 1: Draw a concept map of your own.

You can do this on a sheet of paper or on a computer. If you draw it out, please take a picture of it so you can upload it to AI.

Step 2: Ask the AI to generate a concept map with the same prompt.

Note that it’s still not quite there . . . but it’s getting closer.

Step 3: Compare your concept map to what it creates.

What trends do you notice? What connections do you have that might be different from what it generated?

Step 4: Ask AI to analyze your concept map

Upload your concept map and give the prompt, “I am (grade level) student in a (describe class). We are learning about _______. I have created a concept map. I would like you to analyze it and tell me where I might have some misconceptions. Also let me know what concepts I haven’t included and why I should consider connecting them.”

Notice that this requires students to start with their own concept maps. We are also asking them to think critically about how they have conceptualized the content in comparison to the AI. The goal is a deeper understanding of the concepts.

 

Strategy #4: Using AI as a Socratic Tool

With this approach you ask AI to pose questions to you and then you pose questions to the chatbot so that you can think critically about the content you are learning. As an added bonus, you might ask the chatbot to play the role of three people with differing views and you can discuss it in-depth.

If you’re not familiar with the Socratic Seminar Process, here is a quick video

Here’s an example prompt:

“I am a sophomore in high school learning about the ethics of AI and the environment. Play the role of climate scientist who is worried about AI and water, one who has a more nuanced view, and one who thinks AI will help conserve water. Be sure that each person uses actual facts. So keep the temperature tight. But also be creative, human, and conversational in tone. No one should “speak” twice without me getting to speak or ask a question. It should be a free-flowing conversation. Explain each person and their fictional role at the start.”

Note that it typically starts out as pretty stilted. So, you need to make it more conversational and I have to ask for it to be more like a Socratic Seminar and not a panel discussion. Check out the video to see why we aren’t quite there yet with Socratic Seminars. I also included the idea of Notebook LM. You’ll see that afterward.

But also, here’s another option. I can use a tool like Notebook LM, drop in a few articles, perhaps some notes I’ve scribbled down when listening to a podcast, and let it create a fictional podcast episode about those ideas. If you listen to this example, you’ll notice that it sounds real. It’s almost creepy the way it captures the “uhs” and “ums” and vocal fries of podcast hosts. But then, when I choose to do so, I can interrupt the podcast host and engage in a discussion.

So if we think about deeper learning, this would be a way for students to make sense out of multiple articles. They could listen to the “podcast” episode ahead of time before reading the actual texts. When confused, they could ask clarifying questions. Or you might have them read multiple articles and summarize them with Cornell Notes or even create a sketchnote connecting the ideas. They they can the podcast as a form of summarizing and synthesizing while also letting them clarify misconceptions.

 

Strategy #5: Interview Fictional Historical Figures

With this approach, students get a chance to interview multiple historical figures and ask their key questions. I recommend making it fictional figures because I’m not a fan of students thinking that a figure like Harriet Tubman or Abraham Lincoln would say something that they never said. But this is a great chance to encourage deeper questioning techniques.

Step 1: Choose your historical characters.

Pick two fictional people based on real history. Create a general profile for them. For example, we might have two women in ancient Greece – one from Ancient Sparta and another from Ancient Athens. Make sure they come from different places, backgrounds, or viewpoints so the contrast will lead to deeper thinking. You could also choose two characters who might disagree on a particular issue.

Step 2: Develop deeper character profiles.

Write a few sentences that describe each person. Include details like where they live, what their daily life is like, what they value, and how they see the world. You’ll feed this to the AI so it can “play” the role during the interview.

Step 3: Provide the prompt for your AI

Tell it who you want to speak with by pasting in one of your character descriptions. For example: “I am a _____ student. I am learning about _______. I would like you to pretend to be two women. I will be giving you their profiles in a moment. I would like to interview both of you. I’ll be asking questions and you can each answer me but also react to each other.” Then you can paste your profiles.

Step 4: Start the interview.

Ask an open-ended question like, “What is your role in your community?” or “What do you think of your city’s values?” Let the AI respond in character. Then repeat this step with the second historical figure.

Step 5: Ask follow-up questions.

Dig deeper based on what each person says. For example, if the Spartan woman talks about training, ask, “Did you ever want more freedom outside of your role?” If the Athenian woman mentions family life, ask, “Did you ever feel limited by your responsibilities?”

Step 6: Let the characters respond to each other.

Now that you’ve heard from both characters, ask them what they think of each other’s way of life. You can even have them “talk” directly to one another through the AI. This turns your interview into a dialogue and helps you explore the differences and similarities between them.

Step 7: Reflect on what you’ve learned.

Look over the conversation and think about what surprised you, what questions it raised, and what it taught you about history. What did you learn about culture, identity, or perspective? This is where the real learning lives.

 

Strategy #6: Using AI as a Simulator

I’ve written before about how our students will need to solve “wicked problems” in the future. I always imagine someone saying this in a Boston accent but maybe that’s just me. Wicked problems are essentially problems where the solutions lead to new problems.

You can use tools like ChatGPT to explore messy, real-world issues—problems with no easy answers and lots of trade-offs. These are called “wicked problems.” They’re complex, full of competing interests, and the decisions you make often ripple out in ways you didn’t expect. AI can help you run a simulation where you make tough choices, test out your ideas, and see how things shift over time. The goal is to practice real-world problem-solving by exploring decisions, unintended consequences, and different perspectives.

Ahead of time:

Choose a wicked problem.

Pick something complex and real—something that doesn’t have one clear solution. You might focus on redesigning a city’s transportation system, regulating AI in schools, or addressing food insecurity in a neighborhood. This is where you, as a teacher, can select a problem that connects to your content standards. The following are the directions for students.

Step 1: Set the scene.
Use a prompt like this to get the AI started:
“You are going to simulate a wicked problem. Scenario: [Describe the challenge]. Act as a facilitator. Introduce key stakeholders, present the first challenge, and show how my decisions lead to both intended and unintended consequences. Let’s begin.”
The more detail you give, the more realistic the simulation will feel. At this point, you hand it off to the students. Add some details about yourself and maybe even some articles, data, etc. that would be helpful for the simulation.

Step 2: Meet the stakeholders.

At this point, you can ask the AI to list out the people involved. Who are the 4–6 key players? What do they want? What are they worried about? This helps you step into their shoes and see the problem from different angles.

Step 4: Face your first decision.

Ask the AI to give you the first major choice you’ll need to make. Something like:

“What’s the first big decision here? Give me a few options and explain what might happen next.”

Make your choice, then pay attention to how the system responds.

Step 5: Follow the ripple effects.

After you decide, ask the AI:

“What are the unintended consequences? How do different stakeholders respond?”

This is where the simulation starts to feel real. You’ll see new tensions, surprises, maybe even conflict.

Step 6: Shift and adapt.

Use what you’ve learned to adjust your approach. Ask:

“Based on what just happened, what should we do next? Give me a few new options.”

Repeat this cycle a few more times—each round helps you understand the system a little more deeply.

Step 7: Reflect on the big picture.

When you’re ready to wrap up, ask the AI to help you reflect:

“How did the situation change over time? Who gained or lost influence? What can I learn about solving complex problems?”

This is where things click. You start to see how systems interact, how change happens, and how important it is to keep thinking, questioning, and adjusting.

Strategy #7: Error Analysis and Feedback Loop for Skills

This one is similar to the previous strategy with concepts. However, this focuses on skill practice instead.

Step 1: Submit multiple assignments.

Use the prompt, “I am ______. I am learning about ________. I’m going to submit a few assignments. I would like you to analyze my answers and highlight errors, misconceptions, or missing reasoning. Provide corrective feedback and resources to fix each error.”

Step 2: Summarize the feedback

Look at the feedback and jot down what you notice.

Step 3: Ask for some basic skill practice. Work through the skill practice handout.

Step 4: Use the chatbot to grade the skill practice and explain where you got it right or wrong.

Step 5: Take handwritten notes on key ideas from the informational text and jot down a few questions you have.

Step 6: Ask clarifying questions to build your own background knowledge focused on any part of the information you don’t understand.

Step 7: Ask the AI to test you on what you just learned. An optional prompt might be:

 

Strategy #8: Metacognitive Coaching Companion

Metacognition is critical for deeper learning. It’s what helps students determine what they know, what they don’t know, and what they need to do next. On a deeper level, metacognitive skills help students with success in college or trade schools, with life-long learning, and with success in the workforce. Here’s a quick overview of the metacognition cycle:

So, what does it look like to use AI to coach students through metacognition? Here’s a quick approach.

Step 1: Study or work on your assignment.

Spend focused time reading, solving problems, reviewing notes, or working through a task. Don’t worry about being perfect. Simply stay present and give it your best effort.

Step 2: Journal your reflection.

Right after your study session, answer a few reflection prompts. Be honest. This isn’t graded. It’s just for you. Try questions like:

• What was easy?

• What was challenging?

• What do I know?

• What don’t I know?

• How did I overcome difficulties?

• What do I still feel unsure about?

Step 3: Create a prompt.
Use a tool like ChatGPT and paste your journal entry into the chat. Then ask something like: “I am a ______. I’ve been learning about _______.  I want you to take on the role of a metacognitive coach who can guide me with questions so I improve on my self-reflection. But also give feedback when necessary.”

Step 4: Read the response and make a plan.

Engage in a question and answer coaching session. This should involve a back and forth conversation.

Step 5: Ask for Advice

Use the prompt, “Analyze my journal and suggest one habit or strategy I can try next time to improve based on what you see and based on our conversation.”

Pay attention to the suggestion the AI gives you. It might be about how to stay focused, how to review material, or how to handle tricky concepts. Write it down. Stick with the concept of “ink it to think it.” Make it your game plan for the next session.

Step 5: Set a small goal.

Before your next study session, set one clear, doable goal based on your reflection. For example: “I’ll try summarizing each section in my own words,” or “I’ll pause every 10 minutes and check my understanding.”

Step 6: Repeat the cycle.

After your next study session, reflect again, journal your thoughts, and check in with the AI. Over time, you’ll start to notice patterns in how you learn—and you’ll get better at adjusting your strategies. You can even ask AI for some trends.

Notice how this process blends together student journaling and self-reflection with AI prompting to help with a more externally focused processing. This helps ensure students don’t fall into cognitive atrophy but instead engage in deeper learning while using AI as a thought partner.

 

Strategy #9: Feynman‑Style Teach‑Back

A few months back, Chad Evans from Pennsylvania shared how he has helped teachers use AI for reciprocal teaching, where students start with a theory and AI uses short feedback and questions to guide them in discovery that leads them closer to understanding the theory. But later this had me thinking about a similar concept promoted by famous theoretical physicist Richard Feynman. He believed that true understanding comes when you can simplify an idea clearly and accurately.

Sometimes the best way to understand a concept is to start with your own theory and test it. This strategy is based on the Feynman technique but begins with your own idea first, not the textbook version. In other words, it incorporates elements of reciprocal teaching. The goal isn’t to be right from the start. The goal is to get clearer and more accurate over time.

Step 1: Choose a concept and form your theory.

Pick a topic that makes you curious—something like “how gravity works” or “how the pyramids were built.” Without looking it up, write out your best explanation based on what you already know or believe. Use your own words. It doesn’t need to be perfect. This is a great chance to share your theory with a partner.

Step 2: Explain your theory to the chatbot

Open ChatGPT or another AI tool and paste in your explanation. Use a prompt like:
“I am a _____ student. I am learning about _______. Here’s my theory about [insert topic]. Can you identify what’s accurate, what needs improvement, and what I might be missing? Help me make it more accurate and complete. Give me gentle critiques and questions that lead me toward discovering the accurate theory.”

Step 3: Read the feedback and take notes.

The AI will point out what you got right and where you might need more detail or corrections. Take note of new ideas or terms that come up. You might be surprised at how close—or how far off—you were.

Step 4: Revise your theory.

Based on the feedback, write a stronger version of your explanation. Add missing details. Fix misconceptions. Try to make your theory clearer and more accurate, while still keeping it in your own voice.

Step 5: Repeat the process.

Send your revised version back to the AI for a second round of feedback. You might ask:

“Is this explanation more accurate now? What’s still unclear or incomplete?”

Keep iterating until you feel confident you understand the concept deeply and can explain it clearly.

Step 6: Reflect on your learning.

What changed between your first version and your final one? What surprised you? What helped you shift your thinking? That’s where real learning happens—not just in getting the right answer, but in seeing how your thinking grows.

One version of this would be to have two students working with a single chatbot going through this process together. This adds an extra human component to the process.

Strategy #10: Using AI for Personalized Feedback

This is strategy is based on the 20-minute peer feedback process:

Step 1: Share your work with the AI.
Paste in your writing, project draft, presentation script, or idea. Be specific about what kind of feedback you’re looking for. For example:
“Here’s my script for a podcast episode on renewable energy. Can you give me feedback on clarity, engagement, and structure? But first, can you ask me follow-up questions. Then when I’m ready, I’ll ask for feedback.”

Step 2: Let the AI ask follow-up questions.
Pay attention to any clarifying questions the AI asks. It might want to know your goal, your audience, or what stage of the process you’re in. Take the time to respond. The more context you give, the more helpful the feedback will be.

Step 3: Ask for 5 positives and 1 area to improve.
Once you’ve answered the clarifying questions, ask:
“Give me five specific things that are working well and one area I could improve.”
This keeps the feedback balanced—affirming what’s strong while still pushing you to grow.

Step 4: Summarize the feedback in your own words.
After reading the AI’s response, take a minute to summarize what you heard. Write it out in a sentence or two:
“The feedback suggests that my structure is strong and my tone is engaging, but I need to clarify my main point in the introduction.”

Step 5: Create your next steps.
Based on the feedback, write 1–3 concrete next steps. These could be edits, rewrites, revisions, or reflection questions. For example:

• Rewrite the introduction to make the main idea clearer.

• Add a stronger transition between sections two and three.

• Practice reading it aloud to check the pacing.

Step 6: Repeat as needed.
After making changes, you can go back to the AI for a second round of feedback—or check in with a peer or teacher to get another perspective.

Get the FREE eBook!

Subscribe to my newsletter and get the  A Beginner’s Guide to Artificial Intelligence in the Education. You can also check out other articles, videos, and podcasts in my AI for Education Hub.

Fill out the form below to access the FREE eBook: