Exercises — Chapter 14: Using AI Effectively

Exercise 14.1: Prompt Improvement Workshop (Apply)

Take each of the following weak prompts and rewrite them using the techniques from this chapter. For each rewritten prompt, identify which techniques you used (specificity, context, format instructions, constraints, audience, zero-shot, few-shot, chain-of-thought, role prompting).

Weak Prompt A: "Tell me about climate change."

Weak Prompt B: "Write a cover letter."

Weak Prompt C: "Is social media bad for teenagers?"

Weak Prompt D: "Help me with my statistics homework."

Weak Prompt E: "Explain machine learning."

For each improved prompt, test it on an actual AI tool and compare the outputs. Write a brief paragraph (3–4 sentences) describing how the output quality changed.

Exercise 14.2: The VIBE Check in Practice (Analyze)

Use an AI chatbot to generate a 300-word explanation of one of the following topics:

  • The causes of the French Revolution
  • How vaccines work
  • The economics of minimum wage increases
  • The history of the internet

Then apply the full VIBE Check framework to the output:

  1. V — Verifiable: Fact-check at least three specific claims. Document whether each claim is accurate, inaccurate, or partially accurate. Cite the sources you used to verify.
  2. I — Internally consistent: Read the full response and identify any contradictions or tensions between different parts.
  3. B — Balanced: Does the response present multiple perspectives where relevant? What is missing?
  4. E — Evidence-backed: Which claims are supported by specific evidence, and which are stated as bare assertions?

Write up your findings in 400–500 words.

Exercise 14.3: The Academic Integrity Spectrum (Evaluate)

For each of the following scenarios, (a) place it on the spectrum from Level 1 (research assistance) to Level 5 (wholesale generation) from Section 14.4, (b) explain your reasoning, and (c) describe what Priya's five-question framework would reveal.

  1. Priya asks a chatbot to explain the difference between "correlation" and "causation" because her textbook's explanation confuses her.
  2. Priya asks a chatbot to generate an outline for her political science paper, then writes the paper herself following that outline.
  3. Priya writes a full draft of her paper, then asks a chatbot to "make it sound more academic."
  4. Priya asks a chatbot to write three paragraphs analyzing a Supreme Court case, then paraphrases them into her own words.
  5. Priya asks a chatbot to write her entire paper, reads through it, changes a few sentences, and submits it.
  6. Priya asks a chatbot to generate 20 potential thesis statements for her paper. She chooses the one she likes best and writes the paper around it.
  7. Priya uses an AI tool to check her paper for grammar and clarity, similar to Grammarly.

Exercise 14.4: Professional Scenarios (Evaluate)

For each scenario below, identify (a) the potential benefits of using AI, (b) the potential risks, and (c) the safeguards that should be in place.

Scenario A: A journalist uses an AI tool to summarize 500 pages of court documents related to a corruption investigation. She plans to use the summaries to identify which sections to read in full.

Scenario B: A small business owner uses an AI chatbot to draft a contract for a freelance designer. He plans to send the contract without having a lawyer review it.

Scenario C: A therapist uses an AI tool to generate discussion questions for a group therapy session on grief and loss.

Scenario D: A teacher uses an AI tool to generate a rubric for grading student presentations on climate change.

Scenario E: A financial advisor uses an AI chatbot to research investment options for a client and emails the AI's analysis directly to the client.

Exercise 14.5: Build Your Personal AI Policy (Create)

Using the template from Section 14.6, write your own personal AI use policy. This should be at least 300 words and should address:

  1. Your core values regarding AI use
  2. Specific guidelines for academic contexts (if applicable)
  3. Specific guidelines for professional contexts (current or anticipated)
  4. Your verification commitments
  5. Skills you want to protect from atrophy
  6. Privacy boundaries
  7. A realistic review schedule

After writing your policy, share it with a classmate or colleague. Discuss: Where do your policies agree? Where do they differ? What does the difference reveal about your respective values and circumstances?

Exercise 14.6: The Prompt Engineering Challenge (Apply/Create)

This is a structured exercise designed to build your prompting skills through deliberate practice.

Round 1: Choose a topic you know well. Ask an AI chatbot three versions of the same question: - Version A: A vague, unstructured prompt (1 sentence) - Version B: A moderately specific prompt using 2–3 techniques from Section 14.2 - Version C: A highly specific prompt combining 4+ techniques

Compare the three outputs. Which is most useful? Which is most accurate? Document the comparison.

Round 2: Choose a topic you know nothing about. Repeat the three-version process. This time, add a VIBE Check evaluation of Version C. How does your ability to evaluate output change when you lack domain knowledge? What does this tell you about the limits of AI-assisted research?

Round 3: Ask a chatbot to help you with a real task you actually need to accomplish this week. Apply everything you have learned. Document your prompt, the output, your evaluation, and your revisions. Reflect on the process in 200 words.

Exercise 14.7: Peer AI Audit (Analyze/Evaluate)

Work with a partner. Each of you should use an AI tool to produce a 250-word summary of the same topic (choose from: the causes of World War I, the psychology of decision-making, the pros and cons of universal basic income, or another topic you both agree on).

Then exchange outputs (but not prompts). Each partner should:

  1. Apply the VIBE Check to the other person's AI output
  2. Identify three strengths and three weaknesses
  3. Guess what techniques the other person used in their prompt

After the evaluation, share your original prompts and discuss: - How did prompt differences lead to output differences? - What did the evaluator catch that the prompter missed? - What does this tell you about the value of external review?