COMP 3270 / Module 4 / Unit 12 / Lesson 12.2

Unit 12 · Lesson 12.2 · Week 8

Productive Struggle, Scaffolding, and Feedback

Design AI assistance that preserves productive struggle, provides scaffolding, and offers formative feedback — without replacing your own learning effort. Learn to request graduated hints, evaluate AI-generated feedback, and plan for scaffold fading.

Estimated study time: 3–4 hours Primary reading: Chapter 12 Prerequisites: Lesson 12.1; TRACE & CLAIM

Table of Contents

Lesson Overview

In this second lesson of Unit 12, you will explore how to design and use AI assistance that preserves productive struggle, provides scaffolding, and offers formative feedback — all without replacing your own learning effort. You will learn to distinguish between struggle that strengthens learning and struggle that is unproductive or caused by inaccessible design. You will also develop skills to request graduated hints rather than complete answers, and to evaluate AI-generated feedback using evidence and criteria.

This lesson builds directly on Lesson 12.1: you will apply the concepts of learning goals, assistance, and substitution to design supportive interactions that keep you at the centre of the learning process. The skills you develop here prepare you for Lesson 12.3 (retrieval and reconstruction) and Lesson 12.4 (the LEARN framework and personal workflows).

📘 Learning and academic-integrity reminder: This lesson is formative and ungraded. It focuses on learning strategies, not assessment completion. Always follow your course, assignment, and institutional policies regarding AI use.

Lesson Learning Outcomes

By the end of this lesson, you will be able to:

  1. Explain the concept of productive struggle and its role in learning.
  2. Determine when struggle supports learning and when additional support is needed.
  3. Request graduated hints rather than complete answers from AI.
  4. Use AI for scaffolding, questioning, explanation, comparison, and formative feedback.
  5. Evaluate AI-generated feedback using TRACE, CLAIM, and other criteria.
  6. Preserve learner effort and decision-making when using AI assistance.
  7. Design graduated-hint sequences and scaffolding strategies.
  8. Plan for scaffold fading as learner competence increases.
  9. Identify when to consult a human instructor, tutor, or other support.
  10. Recognise common risks and limitations of AI-generated feedback.

Before You Begin

Retrieval Practice: Units 10 & 11

Before we dive into new material, retrieve what you know about verification and evidence. Attempt each question from memory before opening the answer.

Question 1. What is the difference between a reasoning error and a hallucination in AI-generated content?

Show answer

Answer: A hallucination is fabricated content with no basis in training data; a reasoning error is flawed logic or inference even when facts are correct.

Explanation: Hallucinations are often harder to detect because they sound plausible. Reasoning errors may be identified by examining the logical flow.

Question 2. What is the primary purpose of the CLAIM framework?

Show answer

Answer: CLAIM provides a systematic process to verify claims: identify the claim, locate the source, assess authority, check corroborating evidence, and determine claim strength.

Question 3. Why is it important to verify an AI-generated explanation before using it to learn?

Show answer

Answer: AI-generated content can be incorrect, misleading, or outdated. If you learn from unverified content, you may internalise errors and develop misconceptions.

Question 4. What is one practical strategy you can use to verify a factual claim made by an AI?

Show answer

Answer: Consult a trusted, authoritative source (e.g., textbook, peer-reviewed article) and compare the AI's claim against it.

Question 5. How does recognition differ from recall?

Show answer

Answer: Recognition is identifying information when it is presented; recall is producing information without seeing it.

Explanation: Recognition is easier and can create an illusion of understanding, but recall is a stronger indicator of learning.

Opening Scenario: Two Approaches to Struggle

Scenario

In an interdisciplinary course on climate change policy, students are asked to analyse a complex dataset and propose mitigation strategies. Two students, Alex and Jamie, approach the task differently.

  • Alex feels unsure about the statistical methods. After reading the problem, Alex immediately enters the entire prompt into an AI system and asks for a complete solution. The AI provides a detailed analysis with recommendations. Alex reads it, feels satisfied, and submits it.
  • Jamie also finds the problem difficult. Jamie spends 45 minutes trying to understand the dataset, experimenting with different visualisations, and sketching out possible approaches. After several dead ends, Jamie asks the AI: "What are the key steps to analyse this type of data? I've tried X and Y but got stuck on Z. Can you give me a hint about the next step?" The AI provides a structured hint, not a full solution. Jamie uses it to continue working and eventually produces a unique analysis.

Questions for reflection:

  1. What is unproductive about Alex's approach? What is unproductive about Jamie's approach (if anything)?
  2. What could a graduated-hint sequence look like for this problem?
  3. When should a learner seek human assistance rather than AI?
  4. How should learning progress be checked in Jamie's case?
Show guided analysis

Analysis:

  • Alex: Alex's approach is unproductive because it bypasses productive struggle entirely. The AI performed the analysis, so Alex did not develop the data analysis skills. This is AI-substituted learning. Alex may also be violating academic integrity if the assignment is graded.
  • Jamie: Jamie's approach is productive: they attempted independently, identified specific difficulties, and requested a bounded hint. However, Jamie might have struggled too long without seeking help; a balance is needed. Jamie's approach is a model of AI-supported learning.
  • Graduated-hint sequence: 1) Clarify the type of data and the goal. 2) Suggest possible visualisation techniques. 3) Point to a specific statistical test that might be appropriate. 4) Provide a partial example of how to apply the test. 5) If still stuck, offer a more detailed example, but not the full solution.
  • Human assistance: If Jamie cannot make progress after several hints, or if the AI's hints are not helpful, they should consult the instructor, a tutor, or a librarian.
  • Checking progress: Jamie should periodically test their understanding by explaining the method to someone else, or by applying it to a different dataset.

Key insight: Productive struggle is effortful but directed. Support should be just enough to keep the learner moving forward, not so much that they stop thinking.

Core Content

1. Productive Struggle

Productive struggle is sustained learner effort on an appropriately challenging task that contributes to understanding, strategy development, and problem-solving. It is the cognitive effort you invest when you are working at the edge of your current ability — not so easy that you are bored, and not so hard that you are overwhelmed.

Productive struggle is not:

💡 Key principle: Support should preserve meaningful learner effort. The goal is to keep the learner in the "struggle zone" where learning happens, but with enough support to prevent frustration and disengagement.

Example: A student learning to write a persuasive essay might struggle to structure their argument. Productive struggle means they try several outlines, get feedback, and revise. Unproductive struggle would be staring at a blank page for hours without any strategy or support.

Common misconception: "All struggle is good." — This is false. Struggle that is caused by poor design, lack of prerequisites, or lack of support is not productive.

2. Scaffolding

Scaffolding is temporary and adjustable support that helps a learner perform a task they cannot yet complete independently. The term comes from construction — scaffolding is erected to support workers building a structure, and it is removed once the structure can stand on its own.

Possible scaffolds include:

Effective scaffolding is:

📌 Important: Scaffolding is not the same as providing the answer. Providing the answer removes the need for the learner to think. Scaffolding supports thinking.

Example: A learner is writing a research proposal. A scaffold could be a template with guiding questions for each section. The learner fills in the content, but the structure is provided. As the learner becomes proficient, the template is removed.

3. Graduated Hints

Graduated hints are a sequence of supports that move from less direct to more direct assistance. They are designed to give the learner just enough help to continue, without giving away the entire solution.

A typical graduated-hint sequence might look like this:

  1. Ask the learner to restate the problem — ensures they understand the task.
  2. Identify the relevant concept — points to the key idea needed.
  3. Point to a likely first step — gives a nudge without solving.
  4. Provide a partial structure — offers a framework to fill in.
  5. Demonstrate a related but different example — shows the method in a similar context.
  6. Reveal a solution only when appropriate — for practice, not for assessed work.
⚠️ Do not reveal assessed answers when doing so would violate course requirements. Even if the learner requests it, you should not provide answers that would substitute for their own work.

Example: A student is stuck on a calculus problem. A graduated hint sequence could be: 1) "What is the function asking you to find?" 2) "Which rule would you use to differentiate this?" 3) "Try applying the chain rule — what is the inner function?" 4) "Here is a similar problem worked out, try to follow the steps."

4. Formative Feedback

Formative feedback is information provided to learners during the learning process to support improvement. It is not a final grade or judgment; it is a tool for growth.

Useful formative feedback should be:

AI can generate formative feedback, but you must evaluate and verify it. AI-generated feedback may be incorrect, generic, overconfident, or misaligned with your actual work.

💡 Tip: When using AI for feedback, provide the AI with the learning goal, the rubric (if available), and your specific work. Ask the AI to point to specific evidence in your work. Then verify the feedback against the rubric and your own judgment.

Example: You write a draft of an essay. You ask the AI: "Please provide feedback on my draft, focusing on clarity of argument and use of evidence. Use the following rubric: [rubric]." The AI gives feedback. You then check if the feedback matches the rubric and your own assessment. You revise accordingly.

5. Risks of AI-Generated Feedback

AI-generated feedback is not infallible. Be aware of these common risks:

You must always evaluate AI feedback using TRACE and CLAIM, and cross-check with authoritative sources and human judgment.

6. Scaffold Fading

Scaffold fading is the gradual removal of support as the learner becomes more competent. The goal is to transfer responsibility from the scaffold to the learner.

Effective fading is:

Example: When learning to write a lab report, you might start with a detailed template (high scaffolding). After a few reports, you move to a checklist (medium scaffolding). Finally, you write without any template (no scaffolding).

✅ Key practice: Plan for scaffold fading from the start. Ask yourself: "When will the learner no longer need this support? How will I know?"

7. When to Seek Human Assistance

AI is a tool, not a replacement for human expertise. You should seek human assistance when:

Human support can come from instructors, tutors, librarians, accessibility services, counselling services, and peers. Do not rely on AI for issues that require human judgment or empathy.

8. When to Stop Using AI

You should stop or limit AI use when:

These stop conditions are signs that AI use is becoming substitutive or risky. Be proactive in setting your own boundaries.

Learning-Process Walkthrough: Challenge, Struggle, Scaffolding, Feedback, and Fading

This walkthrough illustrates how a learner moves from encountering a challenge to using scaffolding and feedback, and finally to fading support.

Context: A student is learning to write a literature review in a social sciences course. The learning goal is to synthesise multiple sources and identify themes.
  1. Challenge: The student reads the assignment and feels overwhelmed by the number of sources. They are not sure how to organise the review.
  2. Independent attempt: The student creates a rough outline, grouping sources by topic, but the groups are disjointed and lack a coherent narrative.
  3. Productive struggle: The student spends time trying to connect the groups, but they are stuck. They know the concepts, but the synthesis is not working.
  4. Seeking scaffolding: The student asks the AI: "I'm writing a literature review on the impact of remote work on employee well-being. I have three groups of sources: psychological effects, organisational policies, and technological factors. How can I structure my review to show the interconnections? Please give me a few guiding questions." The AI provides questions like: "How do organisational policies mediate psychological effects? How does technology enable or hinder those policies?"
  5. Using feedback: The student revises the outline based on the guiding questions. They then ask the AI for feedback on the new outline, specifying the rubric. The AI gives feedback: "Your outline now shows connections, but the evidence from sources is not yet integrated. Consider how each source supports or challenges your points."
  6. Verification: The student evaluates the AI feedback against the rubric and their own understanding. They cross-check with an example of a good literature review provided by the instructor.
  7. Reconstruction and fading: The student writes the draft, initially using the outline as a scaffold. As they become more comfortable with the structure, they rely less on the outline and more on their own synthesis.
  8. Human referral: The student is still unsure about the depth of analysis. They consult the instructor during office hours for a final check.
✅ Outcome: The student used AI for targeted scaffolding and feedback, verified the AI's contributions, and maintained ownership of the synthesis. The scaffold faded as the student's competence grew.
Show solution design: Key decisions
  • Learning goal: Write a coherent literature review.
  • Independent attempt: Create a rough outline.
  • Productive struggle: Identify the gap (lack of connections).
  • Scaffolding: Guiding questions from AI to structure the review.
  • Feedback: AI feedback on the outline, verified against rubric.
  • Fading: Reducing reliance on the outline as writing progresses.
  • Human referral: Instructor consultation for deeper guidance.

Worked Example: Designing a Graduated-Hint Sequence

This worked example demonstrates how to design a graduated-hint sequence for a typical learning problem, using a synthetic scenario.

Scenario

A student is learning to calculate the net present value (NPV) of an investment project in a finance course. The learning goal is to apply the NPV formula correctly and interpret the result.

Graduated-Hint Sequence for NPV Problem
Hint Level Hint Text Purpose
1. Clarification "What are the cash flows for each period? List them in a table." Ensure the learner understands the data.
2. Concept cue "What is the formula for NPV? Write it down and identify each variable." Recall the formula and its components.
3. Strategic question "How do you calculate the present value of each cash flow? Which discount rate do you use?" Guide the learner to the next step.
4. Partial structure "Create a table with columns: Period, Cash Flow, Discount Factor (1/(1+r)^n), Present Value. Fill in the first two rows." Provide a framework to work within.
5. Related example "Here is an example with a different project: [example]. Try to follow the same steps for your project." Show the method in a similar context.
6. Full explanation (if appropriate) Only if the learner is still stuck and this is a practice (not assessed) activity. "Here is the full solution. Study it, then try to solve a similar problem on your own." Provide a model for study, but not for submission.
Show analysis and rationale

Analysis: The hint sequence moves from general to specific, from recall to application. Each hint builds on the previous one, and the learner is encouraged to do the work themselves. The final hint (full solution) is only provided if the learner needs it for study, and only in a practice context. This design preserves productive struggle and learner agency.

Key point: The sequence is flexible. If the learner solves it at hint 3, they do not need hints 4–6. The learner's response determines the next step.

Guided Learning Activities

Activity 1: Productive or Unproductive Struggle?

Purpose: Distinguish productive from unproductive struggle. · Time: 15 min.

Instructions: Classify each scenario:

  • (PS) Productive struggle
  • (US) Excessive or unsupported struggle
  • (ID) Inaccessible task design
  • (PA) Premature assistance
  • (AS) Appropriate scaffold
  • (II) Insufficient information
  1. A student spends 2 hours trying to solve a mathematics problem, using different strategies and checking their work, before finally succeeding.
  2. A student attempts to read a dense academic article but cannot understand the jargon and has no glossary or dictionary.
  3. A student asks the AI for the answer to a problem immediately after reading the question.
  4. A student uses a checklist to ensure they have included all required elements in a report.
  5. A student is given a complex task with no clear instructions and no support, and they feel hopeless.
Show classification and explanation
  1. PS.
  2. ID — lack of support makes it unreasonably difficult.
  3. PA — no independent attempt.
  4. AS — provides structure without giving answers.
  5. US — no support leads to frustration.
Activity 2: Graduated-Hint Ladder

Purpose: Create a graduated-hint sequence. · Time: 25 min.

Instructions: Choose a learning task and design a 5- or 6-stage hint sequence following the progression: clarification → concept cue → strategic question → partial structure → related example → fuller explanation (where permitted).

Show solution design (partial example)

Task: Learning to write a thesis statement for an essay.

  1. Clarification: "What is the main argument you want to make in your essay? Write it in one sentence."
  2. Concept cue: "A good thesis statement should be arguable, specific, and supported by evidence."
  3. Strategic question: "What are the key points you will use to support your argument?"
  4. Partial structure: "A thesis statement often follows this structure: [Topic] + [claim] + [reasons]."
  5. Related example: "Here is a thesis statement from a similar essay: [example]."
  6. Fuller explanation (if needed): "Here is a detailed guide on thesis statements with multiple examples."
Activity 3: Scaffold Design

Purpose: Design a scaffold for a learning activity. · Time: 20 min.

Instructions: For one of the following contexts, design a scaffold that supports the learner without giving away the answer.

  • Concept learning (e.g., understanding a new theory)
  • Problem-solving (e.g., solving a physics problem)
  • Writing (e.g., writing a business report)
  • Interpreting a chart (e.g., reading a statistical graph)
  • Professional communication practice (e.g., giving a presentation)
Show sample scaffold (interpreting a chart)

Scaffold: Provide guiding questions:

  1. "What is the title of the chart? What is being measured?"
  2. "Which region has the highest sales? Which has the lowest?"
  3. "What is the range of values? Are there any outliers?"
  4. "What trends do you notice? Are there any seasonal patterns?"
  5. "What might explain these differences? What additional data would you need?"

Why it preserves learner effort: The student must examine the chart, extract information, and draw conclusions.

Activity 4: AI Feedback Audit

Purpose: Evaluate AI-generated feedback for quality and accuracy. · Time: 25 min.

Instructions: Below is a sample AI-generated feedback. Use TRACE and CLAIM, and consider the criteria for useful feedback. Identify: accurate feedback, generic feedback, unsupported criticism, missing issue, inaccessible wording, excessive praise, incorrect recommendation.

Student draft: "Climate change is a big problem. It is caused by greenhouse gases. We need to do something about it."

AI feedback: "This is an excellent paragraph. Your argument is clear and well-supported. The language is sophisticated. To improve, consider adding a specific example. However, your paragraph is already very strong. Keep up the good work!"

Show audit results and explanation
  • Excessive praise: "excellent", "very strong" — the paragraph is basic.
  • Generic feedback: "consider adding a specific example."
  • Missing issue: The AI did not point out the lack of clear argument and evidence.
  • Inaccessible wording: "sophisticated" — not defined.

Conclusion: This feedback is not useful.

Activity 5: Scaffold-Fading Plan

Purpose: Design a plan for gradually reducing support. · Time: 20 min.

Instructions: Design a scaffold-fading plan. Specify the initial scaffold, the intermediate scaffold, the final stage, how you will decide when to fade support, and how you will handle a learner who struggles after fading.

Show sample fading plan (writing a lab report)

Initial scaffold: Detailed template with sections and guiding questions.

Intermediate scaffold: Checklist of required sections and key elements.

Final stage: No template or checklist.

Fading decision: When the learner consistently produces complete reports, move to the checklist; when they consistently meet the checklist, move to independent writing.

Handling struggles: Revert to a higher level of support temporarily, then try fading again.

Exercises

Complete these exercises to reinforce your understanding. Attempt each before reviewing the answer.

Exercise 1: Distinguish Productive Struggle from Unnecessary Difficulty

Task: Classify each scenario as productive struggle (PS), unsupported struggle (US), or inaccessible design (ID).

  1. A student spends 30 minutes trying to recall a formula, then looks it up and continues.
  2. A student is given a complex problem with no prerequisites and no hints, and they give up after 10 minutes.
  3. A student works through a difficult problem, making errors, learning from them, and eventually solving it.
Show answers and explanations
  1. PS: The student attempts recall and uses a strategy to continue.
  2. ID: No prerequisites and no support.
  3. PS: Errors are part of the learning process; the student perseveres.
Exercise 2: Identify Premature Answer Provision

Task: A student asks an AI: "What is the answer to question 5 on my problem set?" The AI responds with the full solution. Is this premature answer provision? Why? What should the AI have done instead?

Show answer and explanation

Answer: Yes, this is premature answer provision. The student did not attempt the problem independently. The AI should have asked the student to first try the problem, or provided a hint sequence, or reminded the student of the course policy.

Exercise 3: Create Graduated Hints

Task: You are learning to write a persuasive email to a supervisor. Design a 5-stage graduated-hint sequence.

Show solution design
  1. Clarification: "What is the purpose of your email? What outcome do you want?"
  2. Concept cue: "A persuasive email has a clear subject line, a polite opening, a logical argument, and a call to action."
  3. Strategic question: "What evidence or reasoning supports your request?"
  4. Partial structure: "Try this structure: 1) Subject, 2) Greeting, 3) State your purpose, 4) Provide rationale, 5) Call to action, 6) Closing."
  5. Related example: "Here is an example of a persuasive email for a different request."
Exercise 4: Design a Scaffold

Task: Design a scaffold to help a student learn to interpret a pie chart showing budget allocation.

Show solution design

Guiding questions:

  1. "What is the pie chart showing? What is the total?"
  2. "Which category has the largest slice? Which has the smallest?"
  3. "What percentage does each category represent?"
  4. "Are there any categories that are surprisingly large or small?"
  5. "What conclusions can you draw about the budget priorities?"
Exercise 5: Audit AI Feedback

Task: Identify at least three problems with this AI feedback: "Your essay is good. The introduction could be better. You need more evidence. Overall, it's fine."

Show identified problems
  • Generic: "good", "could be better", "more evidence."
  • Not actionable: Does not say how to improve.
  • Overly positive: "fine" with no substantive critique.
  • Missing specifics: No reference to the actual content.
Exercise 6: Revise Defective Feedback

Task: Revise the feedback from Exercise 5 to be specific, actionable, and evidence-based. Assume the draft is about climate change policies.

Show revised feedback

Revised feedback: "Your essay introduces the topic of climate change policies, but the introduction could more clearly state your thesis: that carbon pricing is the most effective policy. In the body, you mention the Paris Agreement, but you do not provide specific data on emissions reductions. Consider citing a specific study that quantifies the impact. Also, your conclusion restates the introduction; try to synthesise the key findings. Use the rubric to check that you have addressed all criteria."

Exercise 7: Design Scaffold Fading and Human Referral

Task: For the persuasive email task (Exercise 3), design a scaffold-fading plan and specify when you would seek human feedback.

Show solution design

Scaffold fading:

  • High scaffold: Use the 5-stage hint sequence as a guide.
  • Medium scaffold: Use a checklist of key elements.
  • Low scaffold: Write without any checklist, but review against a mental checklist.

Human referral: Seek feedback from a mentor or supervisor when the email is for a high-stakes situation. Consult a human if you are unsure about tone or protocol.

Homework Problems

Formative · Ungraded · Recommended learning practice.

Problem 1: Productive Struggle Identification

Task: A student is learning to solve quadratic equations. They attempt 5 problems, get the first 3 wrong, but learn from their mistakes and solve the last 2 correctly. Is this productive struggle? Justify.

Show answer

Answer: Yes, this is productive struggle. The student made errors but used them as learning opportunities and eventually succeeded.

Problem 2: Scaffold Selection

Task: A student is struggling to organise a research paper. Which scaffold would be most appropriate: (a) providing a complete outline, (b) providing a template with section headings, or (c) giving a list of topics to cover?

Show answer

Answer: (b) Providing a template with section headings. It gives structure without providing content.

Problem 3: Graduated Hints for a Concept

Task: Write a 4-stage graduated-hint sequence for a student learning the concept of "supply and demand."

Show solution design
  1. Clarification: "What is the relationship between price and quantity demanded?"
  2. Concept cue: "The law of demand says that as price increases, quantity demanded decreases. How does this affect supply?"
  3. Strategic question: "What happens when demand increases? How do suppliers respond?"
  4. Partial structure: "Draw a simple graph with price on the y-axis and quantity on the x-axis. Label the demand curve and supply curve."
Problem 4: Feedback Verification

Task: AI gives feedback: "Your hypothesis is not supported by the data." How would you verify this claim? What steps would you take?

Show answer

Answer: 1) Check the data against the hypothesis. 2) Re-run the analysis yourself. 3) Consult the original sources of the data. 4) Look for alternative interpretations. 5) Discuss with a peer or instructor.

Problem 5: Scaffold-Fading Design

Task: Design a scaffold-fading plan for learning to write a business memo. Include high, medium, and low support stages.

Show solution design

High support: Detailed template with prompts for each section.

Medium support: Checklist of required sections and key points.

Low support: Write from scratch, using a mental checklist.

Problem 6: Accessibility Consideration

Task: You are designing a scaffold for a student with a visual impairment. How would you adapt a visual scaffold (e.g., a diagram) to be accessible?

Show answer

Answer: Provide a text description of the diagram, use a screen-reader-friendly format, offer an audio explanation, and ensure the scaffold can be used without visual cues.

Problem 7: Human Referral Scenario

Task: A student is using AI for feedback on a draft. The AI gives contradictory feedback: "Your argument is weak" and "Your argument is strong." What should the student do?

Show answer

Answer: The student should not rely on the AI. They should seek human feedback from an instructor, tutor, or peer.

Problem 8: Transfer Problem

Task: Apply productive struggle to a professional setting, such as learning a new software tool at work.

Show solution design

Answer: In a professional setting, productive struggle means attempting to use the software, making mistakes, and learning from them. Seek help after making an independent attempt.

Problem 9: Comprehensive Analysis

Task: Analyse the following: A student uses AI to generate a list of questions to consider, then writes their own analysis. They ask AI for feedback, but the feedback is vague. They then revise their analysis and submit it.

Show answer

Answer: The AI was supportive in generating questions (scaffolding). The feedback was vague, so it was less useful. The student could have improved by asking for more specific feedback and providing the rubric.

Problem 10: Stop Condition

Task: A student asks AI for a hint on a problem. The AI provides a hint, but the student still cannot solve it. The student asks for another hint, and the AI gives the full solution. Should the student stop using the AI?

Show answer

Answer: Yes, the student should stop because the AI has crossed from scaffolding to substitution.

Problem 11: Design a Learning Interaction

Task: Design a learning interaction where a student uses AI to learn about the concept of "cognitive load." Include a diagnostic question, three graduated hints, and a self-explanation prompt.

Show solution design

Diagnostic question: "What is cognitive load theory?"

Hint 1: "Think about the different types of cognitive load: intrinsic, extraneous, and germane."

Hint 2: "Intrinsic load is the complexity of the material; extraneous load is the way information is presented; germane load is the effort to create schemas."

Hint 3: "Try to explain the theory in the context of learning a new skill."

Self-explanation prompt: "Explain how cognitive load theory can help you design your own study strategies."

Mini-Project: Scaffolded AI Learning Interaction

Purpose

Design a complete AI learning interaction that incorporates productive struggle, scaffolding, graduated hints, feedback, and fading. Estimated effort: 120–180 minutes.

Task

Select a bounded learning activity and develop a scaffolded AI learning interaction that includes the following 17 elements:

  1. Learning goal
  2. Prior knowledge assumptions
  3. Independent attempt
  4. Diagnostic question
  5. Hint 1
  6. Hint 2
  7. Hint 3
  8. Partial example
  9. Self-explanation prompt
  10. Learner response point
  11. AI feedback criteria
  12. Verification requirement
  13. Accessibility consideration
  14. Scaffold-fading plan
  15. Human referral point
  16. Stop condition
  17. Evidence that learning, rather than copying, occurred
Show staged hints

Hint 1: Start with the learning goal. Be specific about what the learner should be able to do after the interaction.

Hint 2: Design hints that are graduated — each hint progressively more direct, but never give the full answer until the final stage (if at all).

Hint 3: Plan for verification. How will the learner verify the AI's feedback? What sources will they use?

Show solution architecture and partial example

Partial example (concept learning: "confirmation bias"):

  1. Learning goal: Define confirmation bias and identify it in a real-world situation.
  2. Prior knowledge: Basic understanding of cognitive biases.
  3. Independent attempt: Student writes a definition and identifies an example.
  4. Diagnostic question: "What is confirmation bias? Provide an example."
  5. Hint 1: "Think about how you might selectively search for information that supports your beliefs."
  6. Hint 2: "Confirmation bias involves favouring information that confirms your existing beliefs."
  7. Hint 3: "Consider a time when you formed an opinion and only looked for supporting evidence."
  8. Partial example: "If you believe a diet is effective, you might only read success stories."
  9. Self-explanation prompt: "Explain how confirmation bias might affect decision-making in your profession."
  10. Learner response: The student writes their revised definition and example.
  11. AI feedback criteria: Check if the definition includes the key elements.
  12. Verification: Student checks the definition against a textbook.
  13. Accessibility: Provide text and audio options, clear language.
  14. Scaffold fading: Next time, the student will write a definition without hints.
  15. Human referral: If the student cannot identify an example, they will ask a peer or instructor.
  16. Stop condition: If the AI starts generating the definition for the student, stop.
  17. Evidence of learning: The student can explain the concept to someone else and identify it in a new context.
✅ Self-review checklist:
  • Did I clearly define the learning goal?
  • Did I include a diagnostic question and graduated hints?
  • Did I design feedback criteria and a verification process?
  • Did I plan for accessibility and human referral?
  • Did I define stop conditions and evidence of learning?

Self-Check Quiz

Test your understanding with these questions. Attempt each from memory before revealing the answer.

1. What is productive struggle?

Show answer

Answer: Sustained learner effort on an appropriately challenging task that contributes to understanding, strategy development, and problem-solving.

2. Which of the following is not a characteristic of effective scaffolding?

Select one:
Show answer

Answer: C) Permanent. Scaffolding is temporary.

3. What is the first step in a typical graduated-hint sequence?

Select one:
Show answer

Answer: B) Ask the learner to restate the problem.

4. Why should you evaluate AI-generated feedback?

Select one:
Show answer

Answer: B) Because it may be generic, incorrect, or misaligned.

5. What is scaffold fading?

Select one:
Show answer

Answer: B) Gradually reducing support as competence increases.

6. When should a learner seek human assistance instead of AI?

Select one:
Show answer

Answer: B) When they are confused about course expectations.

7. Which of the following is a stop condition for AI use?

Select one:
Show answer

Answer: B) The learner cannot explain the result.

8. What is the primary purpose of formative feedback?

Select one:
Show answer

Answer: B) To support improvement during learning.

9. A student receives AI feedback: "Your analysis is thorough, but your conclusion is weak." What should the student do next?

Select one:
Show answer

Answer: C) Verify the feedback against the rubric and their own work, then revise accordingly.

Assignment Connection

Assignment 5: Verification and Fact-Checking Assignment (weight: 10%)

The verification skills from Assignment 5 are essential when using AI for scaffolding and feedback. You must verify that the hints, explanations, and feedback you receive are accurate and aligned with your learning goals.

Assignment 1: AI Learning and Reflection Journal (weight: 15%)

This lesson provides evidence you can preserve for Assignment 1, including: your graduated-hint sequences, scaffold designs, audits of AI feedback, scaffold-fading plans, and reflections on productive struggle and human referral.

Consider adding: your graduated-hint ladder (Activity 2), your scaffold design (Activity 3), your feedback audit (Activity 4), and your mini-project (Scaffolded AI Learning Interaction).

Assignment 1 Journal Connection

Consider the following prompts for your Assignment 1 journal entry.

📓 Journal prompts
  1. Productive struggle: Recall a time when you struggled productively. What made it productive? What support did you have? How did it feel?
  2. Scaffolding: Describe a time when you used a scaffold (from a teacher, AI, or other) that helped you learn. What made it effective?
  3. AI feedback: Have you ever received AI feedback that was not useful? What was wrong with it? How did you respond?
  4. Human referral: Think of a situation where you needed human help instead of AI. What led you to seek human help? What was the outcome?
Show reflection-planning checklist
  • Did you describe a specific instance of productive or unproductive struggle?
  • Did you analyse the effectiveness of a scaffold?
  • Did you evaluate the quality of AI feedback you have received?
  • Did you reflect on when and why you sought human support?
  • Did you identify any changes you will make to your learning practices?

Key Takeaways

Remember these points
  • Productive struggle is effortful but directed learning that leads to growth. Not all struggle is productive.
  • Scaffolding is temporary, adjustable support that helps learners perform tasks they cannot yet do independently.
  • Graduated hints move from less direct to more direct assistance.
  • Formative feedback should be specific, actionable, evidence-based, and aligned with learning goals. AI-generated feedback must be verified.
  • Feedback risks include incorrectness, genericness, overconfidence, excessive positivity, inconsistency, and misalignment with rubrics.
  • Scaffold fading is the gradual removal of support as competence increases.
  • Human referral is essential when AI cannot help, when policies are unclear, or when professional judgment is needed.
  • Stop conditions help you recognise when AI use is becoming substitutive or risky.
  • Always verify AI-generated hints, explanations, and feedback using TRACE, CLAIM, and authoritative sources.
  • Responsible AI-supported learning requires that you remain the primary agent of your own learning.

Key Terms

Productive struggle
Sustained learner effort on an appropriately challenging task that contributes to understanding, strategy development, and problem-solving.
Scaffolding
Temporary and adjustable support that helps a learner perform a task they cannot yet complete independently.
Graduated hints
A sequence of hints that move from less direct to more direct assistance, designed to preserve learner effort.
Formative feedback
Information provided to learners during the learning process to support improvement, not for grading.
Scaffold fading
The gradual removal of scaffolding as learner competence increases.
Human referral
Seeking assistance from a human instructor, tutor, librarian, or other qualified person when AI is insufficient or prohibited.
Stop condition
A signal that it is time to stop or limit AI use, such as when it begins performing assessed work or the learner cannot explain the result.
Feedback audit
The process of evaluating AI-generated feedback for accuracy, specificity, actionability, and alignment with learning goals.
Inaccessible design
Learning materials or tasks that are unnecessarily difficult due to poor design, lack of prerequisites, or lack of accommodation.
Premature assistance
Providing help before the learner has made a reasonable independent attempt.
Excessive praise (feedback risk)
AI feedback that is overly positive and does not accurately reflect the quality of the work.
Generic feedback
Feedback that is vague and not specific to the learner's work.
Calibration
The process of adjusting support to match the learner's current level of competence.

Looking Ahead

In Lesson 12.3: Retrieval, Metacognition, Reconstruction, and Transfer, you will build on the concepts of scaffolding and feedback to focus on how to solidify learning through active retrieval, independent reconstruction, and application to new contexts. You will learn to distinguish recognition from recall, calibrate your confidence, and manage cognitive offloading.

After completing all four lessons in Unit 12, you will be prepared for Unit 13: Research and Scholarly Practice.

Continue to Lesson 12.3 ← Back to Lesson 12.1 Return to Unit 12 Home

Sources and Further Reading

[INSERT VERIFIED LEARNING-SCIENCE REFERENCE ON PRODUCTIVE STRUGGLE]
[AUTHOR OR SUBJECT-MATTER EXPERT REVIEW REQUIRED]

All sources are authoritative, current, and directly relevant to the learning outcomes of this lesson. Some page numbers and specific references require verification from the published textbook.

👤