Chapter 39: Exercises

Conceptual Understanding

Exercise 1: The Collingridge Dilemma Applied Select an AI capability that is currently emerging but not yet widely deployed (examples: AI agents for legal practice, AI-generated pharmaceutical research, AI-powered autonomous vehicles for public transportation). Apply the Collingridge dilemma: (a) identify the potential harms that are difficult to predict before deployment; (b) identify the governance mechanisms that would be difficult to implement once the technology is entrenched; and (c) recommend proactive governance measures that should be developed now, before widespread deployment.

Exercise 2: Ethical Foresight Methodology "Ethical foresight" involves systematically identifying ethical implications of emerging technologies before they become crises. Design a methodology for ethical foresight in an AI context. Your methodology should address: (a) how to identify emerging AI capabilities for ethical analysis; (b) what kinds of stakeholders should be involved in foresight exercises; (c) what outputs the methodology should produce; and (d) how those outputs should be used by organizations and policymakers.

Exercise 3: Agentic AI Risk Assessment You are the Chief Risk Officer of a mid-sized financial services firm. Your technology team has proposed deploying an AI agent to manage customer communications — reaching out to customers with personalized recommendations, responding to customer inquiries, and escalating complex issues to human advisors. Conduct a risk assessment of this proposal that: (a) identifies the specific risks of an agentic deployment; (b) compares these risks to a non-agentic alternative (a chatbot with limited capabilities); and (c) recommends governance requirements that would need to be met before the agent could be deployed.

Exercise 4: Cognitive Liberty Analysis Nita Farahany argues that cognitive liberty — the right to mental self-determination — is an emerging human rights concept made urgent by AI. Apply this concept to three specific AI applications: (a) emotion AI used in job interviews; (b) neuromarketing using brain-computer interfaces; and (c) predictive policing using behavioral pattern analysis. For each application, identify what cognitive liberty interests are implicated and what governance requirements would protect them.

Exercise 5: Power Concentration Scenarios Develop two contrasting scenarios for the governance of frontier AI in 2030: (a) a scenario in which frontier AI remains an oligopoly controlled by a small number of private companies; and (b) a scenario in which significant public investment and antitrust action has produced a more competitive landscape. For each scenario, identify: the implications for democratic accountability, market competition, and international equity. Which scenario do you consider more desirable, and what governance actions would move toward it?

Case Analysis

Exercise 6: Agentic AI Incident Response An AI agent deployed by your organization to manage vendor communications sends an email to a major supplier that effectively threatens termination of the contract unless the supplier agrees to new terms — terms that were not authorized by anyone in your organization. The supplier responds with legal threats. Conduct an incident post-mortem: (a) trace how this might have occurred; (b) identify what governance failures it reveals; (c) assign accountability; and (d) recommend specific governance changes that would prevent recurrence.

Exercise 7: 2024 Election Deepfake Response You are the communication director for a political campaign in a democratic country. Three days before the election, a video surfaces on social media that appears to show your candidate making racist remarks. The video is AI-generated — you are certain of this because you have the authentic footage from which the deepfake was created — but proving this to the general public in 72 hours is extremely difficult. Develop a rapid response strategy that: (a) addresses the specific incident; (b) works within existing platform policies; and (c) considers the broader implications of your response for democratic discourse.

Exercise 8: AI and Elections — Regulatory Design Design key provisions of legislation specifically governing AI in electoral contexts. Your legislation should address: (a) mandatory disclosure requirements for AI-generated political content; (b) prohibition of specific uses of AI in electoral contexts; (c) enforcement mechanisms and penalties; (d) platform obligations; and (e) international coordination provisions. Consider how your legislation would interact with free speech protections in your jurisdiction.

Exercise 9: Human-AI Dependency Assessment Conduct a personal audit of your own AI use over the past month. For each AI tool you used regularly: (a) identify what task it performed; (b) assess whether you could perform that task without AI support if necessary; (c) identify any skills or capacities that may have atrophied through AI use; and (d) assess whether your use is consistent with the "healthy human-AI relationship" characteristics described in this chapter. Based on your audit, what, if anything, would you change about your AI use?

Exercise 10: Climate and AI Tradeoffs A large technology company is considering training a new frontier AI model. The training run will consume approximately 1,000 MWh of electricity — the equivalent of powering 90 average US homes for a year. The company claims the model will improve climate modeling and accelerate climate science. Conduct an ethical analysis of this tradeoff: (a) how should the direct climate cost of training be weighed against the potential climate benefits of the model's applications? (b) what information would you need to conduct this analysis? (c) what governance requirements should apply to high-energy AI training runs?

Forward-Looking and Reflective Exercises

Exercise 11: 2035 AI Ethics Scenario It is 2035. Write a two-page memo from the perspective of a Chief AI Ethics Officer at a major organization, reviewing the AI ethics challenges of the past ten years. Your memo should: (a) identify two challenges that were anticipated in 2025 and adequately governed; (b) identify two challenges that were not anticipated or not adequately governed; and (c) recommend governance priorities for the next decade. This exercise requires you to reason forward from current trajectories while acknowledging genuine uncertainty.

Exercise 12: Your Ethical Red Line The chapter argues that individual ethical responsibility in AI includes "refusing to build systems that are designed to deceive, discriminate, or harm, even when instructed to do so." Reflect on this claim: (a) identify a specific AI system that you believe would cross an ethical line you would not cross professionally; (b) describe what you would do if instructed to build or deploy such a system in your current or anticipated professional role; and (c) identify what organizational and legal protections currently exist for people who make such refusals.

Exercise 13: Letter to a Future Colleague Write a letter to a colleague who will be working in AI ethics in 2040. Describe: (a) what you understand about AI ethics as of today; (b) what you believe will remain relevant; (c) what you believe will have changed substantially; and (d) what advice you would offer. The letter should reflect genuine engagement with uncertainty rather than confident prediction.

Exercise 14: The Urgency-Hope Balance This chapter argues for both urgency ("the window for establishing good norms is narrowing") and hope ("democratic societies have successfully governed powerful technologies"). Write an analysis that: (a) assesses the evidence for urgency in AI governance; (b) assesses the evidence for the possibility of effective governance; and (c) identifies what conditions or actions would make effective AI governance more likely.

Exercise 15: Policy Brief for a New Government A newly elected government in a democratic country has asked for a policy brief on AI governance priorities. The brief should be two to three pages and address: (a) the three most urgent AI governance challenges requiring near-term action; (b) the one long-term governance investment that will matter most in ten years; and (c) one international coordination priority. Your brief should be realistic about what government can and cannot accomplish.

Design and Application

Exercise 16: AI Governance Framework Design Design an AI governance framework for a specific organization type of your choosing (hospital system, law firm, retail company, university, government agency). Your framework should include: (a) a risk classification system for AI deployments; (b) governance requirements for each risk level; (c) organizational structures and roles; (d) monitoring and review mechanisms; and (e) processes for raising and resolving AI ethics concerns.

Exercise 17: Cognitive Liberty Policy Draft an organizational policy on cognitive AI — AI systems that assess, infer, or respond to employee cognitive and emotional states. Your policy should address: (a) what uses of cognitive AI are permitted; (b) what disclosures are required; (c) what data protections apply; (d) what rights employees have to contest AI-generated assessments; and (e) enforcement mechanisms.

Exercise 18: International AI Governance Institution Design the key features of an international AI governance institution. Address: (a) its mandate and scope; (b) membership and decision-making structure; (c) relationship to existing institutions (UN, OECD, WTO); (d) enforcement mechanisms; and (e) how it would represent the interests of countries currently underrepresented in AI development.

Exercise 19: AI Ethics Culture Assessment Assess the "AI ethics culture" of an organization you are familiar with — your employer, your university, or an organization whose public information you have access to. Apply the framework from Section 9: Does the organization treat AI ethics as compliance or as practice? What evidence supports your assessment? What would genuine ethical AI culture look like in this organization?

Exercise 20: Personal Commitment Document Write a personal commitment document — not a policy or a framework, but a statement of your own values and commitments regarding AI — that you would be willing to share with a future employer or professional community. The document should address: (a) what you will and will not do professionally in AI contexts; (b) what you commit to doing to maintain and develop your ethical judgment; and (c) how you will engage with AI governance and policy beyond your professional role.

Research and Extended Projects

Exercise 21: Agentic AI Governance Survey Survey the governance frameworks that at least three major AI agent vendors publicly provide — in documentation, terms of service, and responsible use policies. Identify: what governance guidance they provide to enterprise customers; what liability they accept for agent actions; what safeguards they build into their systems; and what is absent from their frameworks that the chapter suggests is necessary.

Exercise 22: Electoral AI Governance Comparison Research how at least four countries governed AI in their 2024 elections. Identify: the legal frameworks that applied; the platform policies that were enforced; the civil society responses; and the documented incidents and their outcomes. Assess which country's governance approach was most effective and why.

Exercise 23: AI Concentration Analysis Using publicly available information, document the current market structure of frontier AI development. Identify: the major actors and their market positions; the barriers to entry for new competitors; the regulatory frameworks that apply; and the current state of antitrust attention to AI concentration. Assess whether current market structure poses governance risks of the kind described in Section 4.

Exercise 24: Climate and AI Research Research the current state of knowledge on AI's energy consumption and climate impact. Identify: the most reliable estimates of AI's current and projected energy use; the geographic distribution of that energy use; the sources of energy used; and the evidence for or against the Jevons paradox applying to AI. Assess what governance interventions would have the greatest impact on AI's climate trajectory.

Exercise 25: Book-Length Reflection Having completed this book, write a three-to-five-page reflection that: (a) identifies the three insights from across the book that you find most valuable or most surprising; (b) identifies the question the book raised but did not answer that you find most important; (c) assesses what you would do differently in a professional or civic role as a result of what you have learned; and (d) identifies what you would read or learn next to continue your AI ethics development.