Chapter 24 Self-Assessment Quiz

Learning in the Age of AI: What's Still Worth Knowing When Machines Can Look It Up

Instructions: Take this quiz without looking back at the chapter. The point isn't to get a perfect score — it's to find out what you actually retained versus what you only think you retained. After you finish, check your answers using the key at the end and note which areas need review. That process — predicting, testing, and correcting — is itself a metacognitive exercise.

Section 1: Multiple Choice

Choose the best answer for each question.

1. The "knowledge paradox" described in this chapter refers to the idea that:

a) The more you know, the less you need AI b) AI is only useful for people who already have advanced degrees c) You need to already know things about a topic in order to use AI effectively for learning about that topic d) Knowledge is becoming less valuable in the AI era

2. Which of the following is the BEST example of using AI as a cognitive tool rather than a cognitive replacement?

a) Asking AI to write your essay and then editing the result b) Asking AI to generate practice quiz questions about a topic you've studied, then answering them yourself c) Asking AI to summarize a reading you haven't read yet d) Asking AI to solve a problem set and then copying the answers

3. The chapter describes AI "hallucination" as:

a) When AI develops consciousness and starts having experiences b) When users imagine capabilities that AI doesn't actually have c) When AI generates confident-sounding information that is factually incorrect d) When AI produces outputs that are too creative to be useful

4. The Explain-Before-You-Ask Protocol involves:

a) Having AI explain a concept to you before you attempt to learn it b) Explaining your current understanding and identifying specific gaps before querying an AI c) Asking AI to explain its reasoning before you accept its answer d) Teaching the AI about yourself before asking it questions

5. According to the chapter, prompt engineering is fundamentally a form of:

a) Computer programming b) Creative writing c) Metacognition — knowing what you know and don't know, then asking precise questions d) Information technology management

6. Automation complacency refers to:

a) The tendency for AI systems to become lazy over time b) The tendency to over-trust automated systems and stop monitoring their performance c) The tendency for humans to automate tasks they enjoy d) The complacency of AI developers about safety issues

7. On the AI Learning Ladder described in the chapter, which rung has the HIGHEST learning value?

a) Rung 1: AI as Answer Machine b) Rung 2: AI as First-Pass Explainer c) Rung 3: AI as Explainer (After Your Own Attempt) d) Rung 5: AI as Practice Generator

8. The term "deskilling" in this chapter refers to:

a) Teaching AI to do skilled work b) The loss of human capability that occurs when a task is fully delegated to technology c) The process of breaking complex skills into simpler components d) AI replacing skilled workers in the job market

9. The chapter argues that the skills AI makes MORE necessary (not less) include:

a) Memorizing facts, speed-reading, and fast typing b) Evaluating information, monitoring understanding, transferring knowledge, and making ethical judgments c) Following instructions, copying accurately, and completing routine tasks d) Programming, data entry, and digital literacy

10. Marcus Thompson's key realization about AI use came when he:

a) Got the highest score in his cohort by using AI extensively b) Submitted AI-generated code, got full marks, and realized he had learned nothing from the experience c) Discovered that AI was better at explaining concepts than his professor d) Was caught using AI by his instructor and received a failing grade

Section 2: True/False with Justification

Determine whether each statement is true or false based on the chapter, then write 1-2 sentences explaining your reasoning.

11. True or False: The chapter argues that students should completely avoid using AI tools for learning.

Your justification: ___

12. True or False: According to the chapter, a complete beginner in a subject is well-positioned to benefit from AI explanations because AI can fill their knowledge gaps.

Your justification: ___

13. True or False: The extended mind thesis suggests that using external tools like AI is a form of cheating.

Your justification: ___

14. True or False: The chapter claims that metacognition is the one skill AI cannot perform on your behalf.

Your justification: ___

15. True or False: According to the chapter, using AI to generate practice questions for yourself is an example of AI as a cognitive replacement.

Your justification: ___

Section 3: Short Answer

Answer in 2-5 sentences. Aim for clarity and precision.

16. Explain why the same AI tool (e.g., ChatGPT) can be either a powerful learning aid or a learning hindrance. What factor determines which role it plays?

17. The chapter draws a parallel between GPS navigation and AI-assisted learning. Explain this analogy. What does GPS navigation teach us about the risks of cognitive offloading?

18. Describe the generation effect and explain how AI use can either preserve or destroy it. Give one specific example of each.

19. Why does the chapter argue that metacognition becomes MORE important in the AI era, not less? Explain the reasoning in your own words.

Section 4: Applied Scenario

20. Read the following scenario and answer the questions that follow.

Scenario: Two students are preparing for a biology exam on cellular respiration. Both use an AI chatbot.

Student A (Layla): Types "explain cellular respiration" into the AI. Reads the response carefully. Highlights the key points. Feels confident she understands it. Moves on to the next topic.

Student B (Chen): First, writes down everything he can remember about cellular respiration from memory. He gets the basic glycolysis steps but is confused about the electron transport chain. He then asks the AI: "I understand that glycolysis breaks glucose into pyruvate in the cytoplasm, but I'm confused about what happens to the NADH molecules in the electron transport chain — specifically, how do they contribute to the proton gradient?" After reading the AI's response, he closes the chat and tries to explain the electron transport chain from memory. He gets stuck on one step, goes back to the AI for that specific point, and tries again.

a) Classify each student's AI use on the AI Learning Ladder (specify the rung). Explain your classification.

b) Which student is more likely to perform well on the exam? Explain your reasoning using at least two concepts from this chapter.

c) Identify the specific metacognitive skills that Chen is demonstrating that Layla is not.

d) Layla's approach involves a specific learning trap discussed in Chapter 1. Name it and explain how AI makes it even more dangerous than rereading a textbook.

Answer Key

Section 1: Multiple Choice

1. c) The knowledge paradox is the counterintuitive idea that you need prior knowledge to use AI effectively — because you need knowledge to evaluate whether the AI's output is accurate, to ask precise questions, and to integrate new information with what you already know. Options (a) and (d) mischaracterize the paradox.

2. b) Using AI to generate practice quiz questions is Rung 5 on the AI Learning Ladder — the highest learning value. The AI creates the testing opportunity, but the student does the cognitive work of retrieval. All other options involve AI doing the cognitive work for the student.

3. c) Hallucination refers to AI generating confident-sounding information that is factually incorrect. It's a structural feature of how large language models work (generating statistically likely text) rather than a bug to be fixed. The key danger is that hallucinated content sounds identical in tone to accurate content.

4. b) The Explain-Before-You-Ask Protocol requires you to articulate your current understanding and identify specific knowledge gaps before querying the AI. This activates retrieval practice, improves metacognitive monitoring, and produces more targeted questions.

5. c) The chapter explicitly argues that prompt engineering is metacognition applied to AI interaction. It requires knowing what you know and don't know (monitoring), formulating precise questions (control), evaluating responses (monitoring again), and iterating (the self-regulated learning cycle).

6. b) Automation complacency is the well-documented tendency to over-trust automated systems and stop monitoring their performance. In the learning context, it means stopping your own metacognitive monitoring because "the AI will catch any errors."

7. d) Rung 5 — AI as Practice Generator — has the highest learning value because it uses AI to create retrieval practice opportunities while the learner does all the cognitive work of actually answering the questions.

8. b) Deskilling is the loss of human capability that occurs when a task is consistently delegated to technology. The chapter gives examples from aviation (pilots losing manual flying skills) and medicine (doctors becoming less skilled at unaided diagnosis).

9. b) The chapter identifies metacognitive and evaluative skills — evaluating information, monitoring understanding, transferring knowledge, making ethical judgments — as the skills that become more important. These are the skills AI cannot do for you.

10. b) Marcus's key realization came when he submitted AI-generated code, received full marks, and recognized that he had learned nothing from the experience. The grade was accurate (the code worked) but the learning value was zero.

Section 2: True/False with Justification

11. False. The chapter explicitly advocates for thoughtful, metacognitively informed AI use — not avoidance. It argues for using AI as a cognitive tool (Rungs 3-5 on the Learning Ladder) while avoiding its use as a cognitive replacement (Rungs 1-2). The message is intentionality, not abstinence.

12. False. The knowledge paradox directly contradicts this. Complete beginners lack the prior knowledge needed to evaluate AI outputs, ask precise questions, or identify hallucinations. AI is most dangerous for those who know the least, because they can't tell good information from bad. Beginners need to build foundational knowledge through their own cognitive effort before AI becomes reliably helpful.

13. False. The extended mind thesis (Clark & Chalmers) argues exactly the opposite — that using external tools is a natural extension of cognition, not cheating. The chapter notes, however, that the extended mind thesis has a complication in the learning context: the cognitive work you offload is the cognitive work you don't learn from.

14. True. The chapter explicitly identifies metacognition as "inherently first-person" — requiring a self that is aware of its own cognitive processes. AI can prompt metacognition (by asking "have you checked your understanding?"), but it cannot be metacognitive on your behalf. It cannot feel your illusions of competence, notice your attention drifting, or experience the tip-of-the-tongue sensation.

15. False. Using AI to generate practice questions is Rung 5 on the AI Learning Ladder — the highest learning value, not a cognitive replacement. The AI creates the testing opportunity, but the student does the retrieval, which is where the learning happens.

Section 3: Short Answer (Sample Responses)

16. The same AI tool can be either a learning aid or a learning hindrance because the technology itself is neutral — what matters is how the learner uses it. The determining factor is whether the learner retains the cognitive work that produces learning (struggle, generation, evaluation, retrieval) or delegates that work to the AI. When you use AI after attempting the work yourself, to clarify specific confusions, or to generate practice opportunities, it enhances learning. When you use AI to skip the struggle entirely, it produces the illusion of learning without the substance.

17. GPS navigation is an analogy for AI-assisted learning because both involve cognitive offloading: delegating a cognitive task (navigation or understanding) to an external tool. Research shows that people who navigate with GPS develop weaker spatial memory than those who navigate independently — the tool gets them to the destination but doesn't build the knowledge. Similarly, AI can get you to the "answer" without building the understanding in your head. In both cases, the tool is excellent for productivity (getting there) but poor for learning (knowing the way).

18. The generation effect is the finding that actively generating an answer — even an incorrect one — produces stronger learning than passively receiving the correct answer. AI can preserve the generation effect when a student attempts the problem first and then uses AI to check or refine their answer (generation happened before AI input). AI destroys the generation effect when a student asks AI for the answer first, reading the AI's response without having tried to generate their own — the generation step is eliminated entirely.

19. Metacognition becomes more important in the AI era because AI handles the retrieval-level cognitive tasks (looking up facts, finding information, generating summaries) while leaving the higher-order cognitive tasks (evaluating accuracy, monitoring understanding, making connections, transferring knowledge) to humans. These higher-order tasks are all metacognitive in nature. Additionally, AI's tendency to hallucinate makes critical evaluation essential, and the ease of using AI as a replacement makes the metacognitive discipline of choosing when and how to use it more important than ever.

Section 4: Applied Scenario (Sample Response)

20a. Layla is operating at Rung 2 (AI as First-Pass Explainer) — she went to AI before attempting to understand the material herself and didn't do deep processing afterward. Chen is operating at Rung 3 (AI as Explainer After Own Attempt), with elements of Rung 5 (using the interaction as a self-testing opportunity by trying to explain from memory afterward). Chen engaged the AI only after identifying specific gaps through his own retrieval attempt.

20b. Chen is significantly more likely to perform well. First, his approach preserves the generation effect — by writing down what he knows first, he's strengthening existing memories and exposing specific gaps. Second, his approach includes retrieval practice — both before (writing from memory) and after (trying to explain without the chat) the AI interaction. Layla's approach involves only passive consumption with no retrieval, which creates familiarity without durable understanding.

20c. Chen is demonstrating metacognitive monitoring — he accurately assessed what he knew (glycolysis basics) and what he didn't (electron transport chain specifics), which enabled a precise question. He also demonstrated metacognitive control — after receiving the AI's answer, he tested himself to verify his understanding rather than simply accepting the feeling of comprehension. Layla demonstrated neither.

20d. Layla is experiencing an illusion of competence (Chapter 1). She reads the AI's clear explanation, the material feels familiar, and she interprets that familiarity as understanding. AI makes this trap even more dangerous than rereading a textbook because AI-generated explanations are often clearer, more personalized, and more engaging than textbook prose — which means the feeling of familiarity is even stronger, and the illusion of competence even more convincing.

Scoring Guide

Metacognitive Note: If you used an AI tool to help you answer any of these quiz questions, pause and reflect on what that choice reveals about your current AI use habits. Was it a tool-use (you answered first, then checked) or a replacement-use (you asked AI before trying)? No judgment — just awareness. The awareness is the skill.

End of Chapter 24 Quiz.