Quiz — Chapter 19: Global Perspectives on AI

Multiple Choice

1. Which term describes the phenomenon whereby EU regulations become de facto global standards because companies prefer to build one product that meets the strictest rules?

a) Techno-nationalism b) Digital sovereignty c) The Brussels effect d) Data colonialism

2. The U.S. approach to AI development is best characterized as:

a) State-directed with strict government oversight b) Regulation-first, with innovation constrained by rights-based frameworks c) Innovation-first, with light and patchwork regulation d) Multi-stakeholder governance with binding international commitments

3. The EU AI Act classifies AI applications into risk tiers. Which of the following would be classified as "unacceptable risk" under the Act?

a) A spam filter that automatically sorts email b) A chatbot that does not disclose it is AI c) A social scoring system that ranks citizens based on behavior d) A medical imaging tool used to assist radiologists

4. Which of the following best describes the "compute divide"?

a) The gap between the speed of consumer computers and enterprise servers b) The gap between countries that can afford massive computing infrastructure for AI and those that cannot c) The division between cloud computing and on-premises computing d) The difference between CPU and GPU processing power

5. The concept of "data colonialism" argues that:

a) Developing countries should not collect digital data b) Global data extraction patterns mirror structural features of historical colonialism c) Only countries with colonial histories produce biased AI systems d) Data protection laws are a form of neo-colonialism

6. China's AI development approach differs from the U.S. approach primarily in that:

a) China invests more private venture capital in AI startups b) China relies primarily on open-source AI models c) China's government plays a more direct, coordinating role in AI development d) China focuses exclusively on consumer AI applications

7. Why does the chapter argue that framing U.S.-China AI relations as a "race" can be problematic?

a) Because China is not actually developing AI b) Because the U.S. has already won the race c) Because it implies a single finish line and can be used to justify shortcuts on safety and regulation d) Because only the EU has a coherent AI strategy

8. Digital sovereignty refers to:

a) The idea that all technology companies should be government-owned b) The principle that nations should control the data, infrastructure, and AI systems affecting their citizens c) The belief that digital technology should replace sovereign governments d) The process of digitizing government records

9. Research on automated content moderation across languages has found that:

a) AI moderation systems perform equally well across all languages b) Systems perform best in English and significantly worse in many other languages c) Non-English languages are easier for AI to moderate because they have simpler grammar d) Content moderation AI does not use language-specific training data

10. Which statement best characterizes the current state of global AI governance?

a) A comprehensive binding international treaty governs AI worldwide b) The UN has full enforcement authority over AI systems c) Global AI governance is in its infancy, with multiple overlapping efforts but no comprehensive mechanism d) Industry self-regulation has proven sufficient for global AI governance

True or False

11. The EU AI Act takes an innovation-first approach, prioritizing market freedom over rights protection. True / False

12. Data localization laws can be used both to protect citizens' privacy and to enable government surveillance. True / False

13. The Global South is exclusively a consumer of AI technology, with no significant AI innovation of its own. True / False

14. The OECD AI Principles have been endorsed by over 40 countries and have influenced national AI strategies worldwide. True / False

15. Techno-nationalism is a new phenomenon that emerged only with the development of artificial intelligence. True / False

Short Answer

16. In 2–3 sentences, explain one reason why a country in the Global South might adopt data localization laws, and one risk of doing so.

17. Describe one way that an AI content moderation system designed in the United States might fail when deployed in a culturally different context. Be specific about the type of failure you are describing.

18. The chapter discusses three major approaches to AI governance (U.S., China, EU). In your view, which approach best protects the interests of ordinary citizens? Explain your reasoning in 3–4 sentences.

Scenario-Based

19. A mid-sized African nation is deciding whether to accept a Chinese company's offer to build a comprehensive "smart city" AI infrastructure at a heavily subsidized price, or to invest in building its own domestic AI capabilities — a slower and more expensive path. Using concepts from this chapter, advise the nation's decision-makers. What factors should they weigh? What questions should they ask? (Answer in 150–200 words.)

20. An international organization proposes creating a "Global AI Safety Board" modeled on the International Atomic Energy Agency (IAEA), with the power to inspect AI systems, set binding safety standards, and sanction non-compliant countries. Evaluate this proposal. What are its strengths? What are the practical obstacles? Would you support it? (Answer in 150–200 words.)

Answer Key

  1. c
  2. c
  3. c
  4. b
  5. b
  6. c
  7. c
  8. b
  9. b
  10. c
  11. False — The EU AI Act takes a regulation-first, rights-based approach.
  12. True
  13. False — The Global South is also a site of genuine AI innovation.
  14. True
  15. False — Techno-nationalism has historical precedents like the Space Race; AI has intensified it.
  16. Sample: A country might adopt data localization laws to ensure that citizen data remains under domestic legal jurisdiction and is protected from foreign surveillance or exploitation. However, a risk is that authoritarian governments may use data localization as a tool to control information flows and monitor their own citizens.
  17. Sample: A content moderation system trained primarily on English-language data might fail to detect hate speech in Amharic or Burmese because it lacks sufficient training data in those languages, while simultaneously over-flagging legitimate political discussion that uses terms or rhetorical styles unfamiliar to the system's training context.
  18. Open-ended; strong answers will reference specific features of at least one approach and connect them to citizen impacts.
  19. Open-ended; strong answers will balance sovereignty, cost, capacity-building, privacy, and geopolitical considerations.
  20. Open-ended; strong answers will engage with both the appeal of the model and practical challenges like enforcement, sovereignty concerns, and the difference between nuclear and AI technology.