Exercises — Chapter 1: What Is Artificial Intelligence?
Part A: Conceptual Questions
A1. ⭐ Define artificial intelligence in your own words without using the words "smart," "think," or "brain."
A2. ⭐ List five AI systems you interacted with in the past 24 hours. For each, briefly describe the task the AI performs.
A3. ⭐⭐ Explain the difference between narrow AI and general AI. Give one example of each — one that exists today and one that is hypothetical.
A4. ⭐⭐ A friend says, "Siri is basically general AI because it can answer any question." How would you respond? Be specific about what Siri can and cannot do.
A5. ⭐⭐ In your own words, explain the AI effect. Why does it matter for public understanding of AI?
A6. ⭐⭐ The chapter uses the analogy of a mixing board (multiple sliders) to describe intelligence. Come up with your own analogy for why intelligence is not a single, measurable quantity. Explain why your analogy works.
A7. ⭐⭐⭐ The 1956 Dartmouth proposal claimed that "every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it." Do you agree with this claim? Why or why not? Identify at least one aspect of human intelligence that you think would be especially difficult to "precisely describe."
A8. ⭐ What does the chapter mean when it says AI literacy is a "civic skill"? Give one example of a civic decision that requires understanding AI.
Part B: Applied Analysis
B1. ⭐⭐ Apply the FACTS Framework to a voice assistant (e.g., Siri, Alexa, Google Assistant). Answer all five questions.
| Letter | Question | Your Answer |
|---|---|---|
| F | What specific task does this system perform? | |
| A | How well does it work, and for whom? | |
| C | Who benefits and who might be harmed? | |
| T | What data was it trained on, and who curated it? | |
| S | Who is responsible when it goes wrong? |
B2. ⭐⭐ A university announces it will use AI to detect plagiarism in student essays. Identify at least three stakeholders (people or groups affected) and describe how each might view this system differently.
B3. ⭐⭐⭐ You read a news article with the headline: "AI Outperforms Doctors in Diagnosing Skin Cancer." Using the FACTS Framework, write three specific questions you would want answered before accepting this claim at face value.
B4. ⭐⭐ Choose one of the four anchor systems (ContentGuard, MedAssist AI, Priya's Semester, CityScope Predict). Explain whether it is an example of narrow AI or general AI, and describe the boundaries of what it can and cannot do.
B5. ⭐⭐⭐ A ride-sharing company uses AI to set dynamic pricing ("surge pricing"). Apply the "C" (Consequences) question from the FACTS Framework in depth: Who benefits and who might be harmed? Consider at least three different groups of people.
B6. ⭐⭐ Find a product or service you use that advertises itself as using "AI" or "machine learning." Evaluate whether the marketing language accurately describes what the system does or whether it overstates the technology's capabilities.
Part C: Research Design & Critical Thinking
C1. ⭐⭐⭐ The chapter argues that the AI effect — redefining intelligence to exclude what machines can do — distorts public understanding. Design a simple survey (5–7 questions) that could test whether this effect operates among your classmates. What would you ask, and how would you interpret the results?
C2. ⭐⭐⭐ Imagine you are on a city council considering whether to adopt CityScope Predict. What information would you need before voting? List at least five specific data points, reports, or testimonies you would request, and explain why each is important.
C3. ⭐⭐⭐⭐ A tech company claims its hiring AI is "bias-free" because it does not use candidates' names or demographic information. Using concepts from this chapter, explain why this claim might be misleading. What other sources of bias could exist in the system?
C4. ⭐⭐⭐ Consider Priya's situation from section 1.5. Draft a one-paragraph university policy on student use of generative AI tools that you think would be fair, clear, and enforceable. Then identify one weakness in your own policy.
Part D: Synthesis
D1. ⭐⭐⭐ Write a 300-word explanation of artificial intelligence for a specific audience: choose either (a) your grandparent, (b) a ten-year-old, or (c) a local politician. Adjust your language, examples, and emphasis for the audience you choose.
D2. ⭐⭐⭐⭐ The chapter introduces four anchor AI systems across very different domains (social media, healthcare, education, criminal justice). Identify one concern that is common to all four systems. Explain why this concern appears across such different contexts and what it reveals about AI as a category of technology.
D3. ⭐⭐⭐ The chapter distinguishes between AI capability (what a system can do) and AI understanding (whether the system comprehends what it is doing). Why does this distinction matter? Give a concrete example of a situation where this distinction has practical consequences.
D4. ⭐⭐⭐⭐ Revisit the Edsger Dijkstra quote that opens the chapter: "The question of whether a computer can think is no more interesting than the question of whether a submarine can swim." What is Dijkstra's argument? Do you agree or disagree? Write a 200-word response that engages with the analogy directly.
Part M: Mixed Practice
M1. ⭐⭐ Match each scenario to the most relevant FACTS Framework question (F, A, C, T, or S). Then briefly explain your reasoning for each.
| Scenario | FACTS Letter | Reasoning |
|---|---|---|
| A hiring tool rejects more female candidates than male candidates for engineering roles | ||
| A hospital cannot determine who is liable when an AI-assisted misdiagnosis leads to patient harm | ||
| A company describes its email sorting tool as "AI-powered decision intelligence" | ||
| A facial recognition system was trained exclusively on photographs of lighter-skinned individuals |
M2. ⭐⭐⭐ For each of the following statements, determine whether it describes narrow AI, general AI, or neither. Justify your answer in one sentence.
- A system that can write poetry, play chess, drive a car, diagnose diseases, and learn new skills it was never trained for.
- A system that generates realistic images from text prompts but cannot have a conversation.
- A calculator app on your phone.
- A customer service chatbot that can handle billing questions but redirects all other queries to a human agent.
M3. ⭐⭐ A news article states: "Artificial intelligence has finally learned to be creative." Using the concepts from sections 1.2 and 1.4, write three critical questions you would ask about this claim.
M4. ⭐⭐⭐ Rank the four anchor systems (ContentGuard, MedAssist AI, Priya's Semester, CityScope Predict) from "most potential benefit" to "most potential risk." Explain your ranking. Then acknowledge one strong counterargument to your ranking order.
Part E: Research & Extension
E1. ⭐⭐⭐⭐ Research a real-world case where an AI system produced biased or harmful outcomes (e.g., Amazon's hiring tool, COMPAS recidivism prediction, healthcare allocation algorithms). Write a 500-word summary that includes: (a) what the system was designed to do, (b) what went wrong, (c) who was harmed, and (d) what the case reveals about the limits of narrow AI. Cite your sources.
E2. ⭐⭐⭐ Find and read one of the sources listed in the Further Reading for this chapter. Write a 250-word response that connects one idea from the reading to one concept from the chapter. Where do the source and the chapter agree? Where do they diverge or add nuance?
E3. ⭐⭐⭐⭐ The chapter mentions that expert predictions for when (or whether) general AI will be achieved vary wildly. Research at least three expert opinions on AGI timelines (from named researchers, not anonymous internet sources). Summarize each position in 2–3 sentences and identify the key assumptions underlying their predictions.