Chapter 38 Exercises: AI, Society, and the Future of Work

Section A: Task Automation Analysis (Exercises 1--5)

Exercise 1: Applying the Tasks Framework

For each of the following occupations, identify three to five core tasks. Classify each task along the routine/non-routine and cognitive/manual dimensions. Then assess each task's susceptibility to AI automation (low, medium, or high) and explain your reasoning.

(a) Commercial real estate broker

(b) Emergency room nurse

(c) Corporate tax accountant

(d) High school English teacher

(e) Restaurant general manager

Exercise 2: The Judgment Dimension

The chapter introduces the "judgment dimension" --- the degree to which a task requires evaluating trade-offs among competing values where the criteria for "good" are themselves contested.

For each scenario below, assess where it falls on the judgment dimension (low, medium, or high) and explain what makes it resistant to or susceptible to AI automation.

(a) A hospital administrator deciding whether to allocate a scarce ICU bed to a 75-year-old patient with a 40% survival probability or a 30-year-old patient with a 70% survival probability who arrived 20 minutes later.

(b) A supply chain manager selecting between two shipping routes: one that is 15% cheaper but uses a carrier with a history of labor violations, and one that is more expensive but uses a carrier with strong labor practices.

(c) A portfolio manager deciding whether to sell a stock that has underperformed for three quarters but is held by a long-standing client who has an emotional attachment to the company.

(d) A content moderation team deciding whether to remove a social media post that is factually accurate but could incite panic during a public health emergency.

Exercise 3: Augmentation Design

You are the Chief People Officer at a mid-size insurance company. The CEO has approved a project to deploy an AI system that can process routine insurance claims (clear-cut cases with standard documentation) in approximately 20 percent of the time a human adjuster requires. Currently, your claims department has 200 adjusters.

(a) Describe two different deployment strategies: one that maximizes cost reduction (the "automation" approach) and one that maximizes human-AI complementarity (the "augmentation" approach). Be specific about how each strategy would change the adjusters' roles, the department's headcount, and the customer experience.

(b) For each strategy, identify three risks and three potential benefits.

(c) The CEO asks you to recommend one strategy. Write a one-page memo making your case, including financial considerations, workforce impact, customer experience implications, and long-term strategic positioning.

(d) How would your recommendation change if the company were: (i) a publicly traded company under pressure to improve quarterly earnings, (ii) a mutual insurance company owned by its policyholders, or (iii) a startup trying to disrupt the insurance industry?

Exercise 4: Predicting Task Displacement

The Eloundou et al. (2023) study found that higher-wage workers with more education were more exposed to LLM automation than lower-wage workers --- the opposite of previous automation patterns.

(a) Explain why this inversion occurs. What is it about the tasks performed by higher-wage knowledge workers that makes them susceptible to LLM capabilities?

(b) The study distinguishes between "exposure" and "vulnerability." Why are higher-wage workers potentially less vulnerable to displacement even though they are more exposed? Identify at least three factors that reduce vulnerability.

(c) Consider a mid-career corporate lawyer (15 years of experience, partner track). Which of their tasks are most exposed to LLM automation? Which are least exposed? How should they adjust their professional development strategy in response?

Exercise 5: Automation vs. Augmentation at Athena

Review the Athena workforce impact data from the chapter:

(a) Calculate the cost per avoided layoff, assuming the $4 million reskilling investment was the primary cost of choosing attrition over immediate layoffs.

(b) Research suggests the average cost of an involuntary layoff (severance, outplacement, legal risk, institutional knowledge loss, remaining employee morale impact) is $15,000--$25,000 per affected employee. Compare the cost of Athena's approach to a layoff-based approach. Which is more expensive?

(c) Athena's approach took 18 months. A layoff-based approach could have been implemented in 3 months. What is the economic value of the 15-month delay? Consider both costs (slower efficiency gains) and benefits (maintained morale, preserved knowledge, lower recruitment costs).

(d) NK asks: "What about the 200 positions eliminated through attrition --- the jobs that were never refilled?" Is attrition truly a "no harm" approach? Who bears the cost? Write a paragraph arguing that attrition is ethically equivalent to layoffs, and a paragraph arguing that it is meaningfully different.

Section B: Workforce Transition Planning (Exercises 6--10)

Exercise 6: Reskilling Program Design

You are designing a reskilling program for 150 customer service representatives whose roles will be significantly changed by AI deployment. Currently, they handle a mix of routine inquiries (60% of volume) and complex cases (40%). The AI system will handle all routine inquiries, leaving the agents to focus on complex cases --- but the complex cases require different skills (emotional intelligence, problem-solving, technical product knowledge) than routine inquiries.

(a) Design a 12-week reskilling curriculum. Specify the weekly topics, the learning methods (classroom, on-the-job, mentoring, simulation), and the assessment approach.

(b) Not all 150 agents will successfully transition. Based on Athena's 78% completion rate, approximately 33 agents may not complete the program. What support should be provided to these individuals?

(c) How will you measure the success of the reskilling program? Identify five metrics, specifying the measurement method and the target value for each.

(d) The CFO asks whether the reskilling program is worth the investment or whether it would be cheaper to lay off the current agents and hire new agents with the desired skills. Prepare a cost-benefit analysis.

Exercise 7: Workforce Impact Assessment

You are the VP of Operations at a logistics company with 5,000 employees. The company is planning to deploy AI systems across three functions: route optimization (affecting 800 delivery drivers), demand forecasting (affecting 40 planning analysts), and automated customer communication (affecting 120 customer service agents).

(a) For each function, analyze: which specific tasks will be automated, which will be augmented, and which will be unaffected. Estimate the net employment impact for each function.

(b) Develop a 24-month workforce transition plan that specifies: timing of deployments, communication strategy, reskilling investments, redeployment opportunities, and support for workers who leave the organization.

(c) The CEO wants to announce the AI deployment as a "workforce empowerment" initiative. Lena Park warns: "Don't use the language of augmentation to sell a strategy of automation." Under what conditions is the CEO's framing honest? Under what conditions is it misleading?

Exercise 8: The Geographic Dimension

A major insurance company headquartered in New York City is deploying AI that will automate 60% of the work performed at its claims processing center in a mid-size city in Ohio. The center employs 500 people and is the largest private employer in the city.

(a) Analyze the potential economic impact on the city if 300 of the 500 positions are eliminated. Consider direct employment effects, multiplier effects on the local economy, tax revenue implications, and social consequences.

(b) What obligations, if any, does the insurance company have to the city beyond its legal requirements? Develop an argument from the stakeholder capitalism perspective and a counterargument from the shareholder primacy perspective.

(c) Propose a transition plan that balances the company's legitimate interest in efficiency with its impact on the community. Consider options such as phased implementation, retraining, severance, community investment, and facility repurposing.

Exercise 9: Global Supply Chain Labor Impact

A US-based technology company outsources AI data labeling to a contractor in the Philippines. The contractor employs 2,000 workers at $4 per hour --- well above the local minimum wage but far below what equivalent work would cost in the US. The company is now developing an AI system that can automate 70% of the labeling work.

(a) What are the company's ethical obligations to the contractor's workers? How do those obligations differ from its obligations to its own employees?

(b) The automation will save the company approximately $8 million per year. Propose an allocation of those savings that balances shareholder returns, investment in AI development, and support for affected workers.

(c) How does this scenario illustrate the "global North-South dynamics" discussed in the chapter? What role should multinational corporations play in managing the global distributional effects of AI?

Exercise 10: Stakeholder Communication

Draft three different communications about the same AI deployment:

Scenario: Your company (a 3,000-employee financial services firm) is deploying an AI system that will automate 40% of the work performed by your 200-person back-office processing team. You expect to reduce the team to 140 people over 12 months, with 30 people redeployed to other roles and 30 positions eliminated through attrition.

(a) A memo to the affected employees, following the chapter's "Principle 1: Be Honest About the Impact."

(b) A press release for external audiences, balancing transparency with legitimate concerns about competitive positioning and market perception.

(c) A board presentation slide deck (outline 5--7 slides) that frames the deployment in terms of strategic value, financial impact, and workforce transition.

(d) Compare the three communications. Where do they differ? Where are you tempted to shade the truth for a particular audience? What are the risks of inconsistency between internal and external messaging?

Section C: Policy Design (Exercises 11--13)

Exercise 11: Lena Park's Governance Triangle

Lena Park's "AI Governance Triangle" identifies three dimensions that any governance system must balance: Innovation, Protection, and Participation.

(a) For each of the following governance approaches, assess its strength on each dimension (strong, moderate, or weak) and justify your assessment: - The EU AI Act - The US "sector-specific" regulatory approach - China's AI regulation framework - Industry self-regulation (e.g., the Partnership on AI) - Open-source AI development norms

(b) Design your own governance framework that you believe achieves the best available balance among the three dimensions. Specify: institutional structure, key rules, enforcement mechanisms, and mechanisms for public participation.

(c) Lena acknowledges that the three dimensions are "in genuine tension." Provide a specific example of a policy decision where increasing one dimension necessarily decreases another.

Exercise 12: UBI vs. Reskilling

The chapter presents two approaches to managing AI's impact on employment: income support (UBI) and skills support (reskilling programs).

(a) Construct the strongest possible argument for UBI as the primary response to AI-driven displacement. Address the three most common objections (cost, work incentives, and treating symptoms rather than causes).

(b) Construct the strongest possible argument for reskilling as the primary response. Address the evidence that retraining programs produce "modest" effects and the challenge of transitioning older workers.

(c) Are these approaches mutually exclusive, or could they be combined? Design a hybrid policy that incorporates elements of both, specifying funding mechanisms, eligibility criteria, and expected outcomes.

(d) How does the answer change depending on the speed of AI-driven displacement? If displacement is gradual (over 20 years), which approach is more appropriate? If displacement is rapid (over 5 years)?

Exercise 13: Corporate Responsibility for AI Transitions

Write a 1,500-word policy paper arguing either:

(a) That corporations have a moral and economic obligation to invest in workforce transition when they deploy AI that displaces workers, and that this obligation should be legally mandated (through transition taxes, mandatory reskilling spending, or required notice periods).

or

(b) That mandating corporate responsibility for workforce transition would slow AI adoption, harm competitiveness, and ultimately reduce the economic growth that creates new employment opportunities. The appropriate response is public investment in education and social safety nets, funded through general taxation.

Your paper should cite specific evidence from the chapter and real-world examples. It should address the strongest counterarguments to your position.

Section D: Ethical Deliberation (Exercises 14--18)

Exercise 14: Tom's Discomfort

Tom says: "I'm on the right side of this disruption. Not everyone is." This reflects a growing awareness among technical professionals that they benefit from the same forces that displace others.

(a) Does Tom have any personal obligation to act on this awareness? If so, what form should that action take?

(b) Is there a meaningful difference between benefiting from AI disruption and causing it? Tom did not automate anyone's job directly. But his skills and career trajectory are made more valuable by the same technologies that make others' skills less valuable. Analyze this distinction.

(c) How should business schools address this dynamic? Should MBA curricula include explicit content on the distributive justice implications of technology leadership? Make an argument for or against.

Exercise 15: The Attrition Question

Athena eliminated 200 positions through attrition --- declining to fill vacancies as employees departed voluntarily. The chapter presents this as a more humane approach than layoffs.

(a) Identify three ways in which attrition imposes costs on workers or communities that are not captured in the characterization of attrition as "no layoffs."

(b) A labor economist argues: "Attrition is just slow-motion layoffs with better PR. The same number of jobs disappear; the only difference is that the company avoids the political and reputational cost of announcing layoffs." Evaluate this argument.

(c) Under what conditions is attrition a genuinely ethical approach to workforce reduction? Under what conditions is it merely a more palatable form of the same outcome?

Exercise 16: The Two Letters

Professor Okonkwo opens the chapter with two letters: one from 1830, one from the present day.

(a) What is the rhetorical purpose of juxtaposing these two letters? What emotional response is Professor Okonkwo trying to evoke, and why is that response pedagogically valuable?

(b) Are the two situations genuinely analogous, or does the analogy obscure important differences between the Industrial Revolution and the AI revolution? Identify three parallels and three differences.

(c) The textile worker writes: "If the machine is to do the work of ten men, then the fruits of that machine must feed ten families, not one." How would you translate this principle into a concrete policy proposal for the AI era?

Exercise 17: The Customer Service Representative's Letter

The second letter Professor Okonkwo reads describes a customer service representative whose team was halved after an AI chatbot deployment. "They told us AI would help us do our jobs better. Then they eliminated our jobs."

(a) Was the employer's initial framing ("AI will help you do your jobs better") necessarily dishonest? Is it possible that the employer genuinely intended augmentation and later shifted to automation? What factors might drive such a shift?

(b) The chapter's Principle 1 states: "Do not tell employees that AI will only 'help them do their jobs better' if you are also planning to reduce headcount." But what if the headcount reduction is not planned at the time of deployment but emerges as a possibility once the AI's capabilities become clear? How should leaders communicate about uncertain futures?

(c) Write a communication that an employer could have sent to the customer service team before the AI deployment that would have been honest about the uncertainty without causing unnecessary panic.

Exercise 18: Building It for Whom?

NK closes the chapter with the question: "We spent thirty-seven chapters learning how to build AI. Now we need to ask: build it for whom?"

(a) Who are the primary beneficiaries of AI as it is currently being developed and deployed? Who bears the primary costs? Are these the same groups?

(b) Could AI be developed differently --- with different priorities, different funding structures, different governance --- such that its benefits were more broadly distributed? Propose three specific changes to the current AI development ecosystem that would shift the distribution of benefits.

(c) Is the question "build it for whom?" sufficient, or should we also ask "build it with whom?" What changes when we include affected communities in the design process, not just the benefit analysis?

(d) Write a personal reflection (500--750 words) on how this chapter has changed --- or confirmed --- your thinking about your role as a future business leader in the AI era. Be specific about at least one decision you would make differently based on what you have learned.