Analyze an Economic Scenario
Economics is not a spectator sport. The frameworks this module has built — general-purpose technology dynamics, productivity channels, the task-based model of automation, displacement and creation mechanisms, inequality analysis, and the policy toolkit — are only as valuable as your ability to apply them to real situations. This lesson gives you a rich scenario and asks you to work through it systematically, producing an economic analysis that a real policymaker or firm leader could use.
The Scenario: AI in a Regional Hospital System
Scenario: A regional hospital system in a mid-sized U.S. city (population 400,000) serves a moderately diverse population with a mix of insured and uninsured patients. The hospital is one of the largest employers in the region, with 6,800 employees across clinical, administrative, and technical roles. The hospital has contracted with a health-tech AI company to deploy three systems simultaneously over the next two years: System 1 — Administrative AI: Automates prior authorization requests, insurance billing code assignment, and patient scheduling optimization. Currently performed by 340 administrative staff. The vendor estimates 70 percent of these tasks can be automated, with one human supervisor per AI system needed for quality assurance. System 2 — Diagnostic AI: An AI tool that reads chest X-rays, CT scans, and mammograms, providing a first-pass analysis before the radiologist reviews. Currently, 18 radiologists process imaging volume that would require 24 without the AI assist. The AI improves diagnostic catch rates for early-stage lung cancer by 12 percent in clinical trials. System 3 — Clinical Documentation AI: An ambient microphone-based AI that generates clinical notes during patient encounters, currently requiring 15-20 minutes of physician time per visit. The AI draft reduces this to 3-4 minutes for review and sign-off. The hospital has 280 primary care physicians and specialists. The hospital's board has asked for an economic analysis before final approval.
A complete economic analysis of an AI deployment examines at least five dimensions: (1) productivity and cost effects, (2) labor-market effects on existing workers, (3) effects on the quality or availability of the service, (4) distributional effects — who gains and who loses, and (5) policy or governance implications. A one-dimensional analysis (this saves money, or this destroys jobs) is always incomplete.
Hospital AI Deployment: Economic Analysis
- Working individually or in groups of 2-3, complete a structured economic analysis of the hospital scenario above. Your analysis should address each of the following questions with specific reasoning drawn from module frameworks.
- PART 1 — PRODUCTIVITY AND GROWTH (Lessons 1-2)
- A. For each of the three AI systems, identify the primary productivity channel: Is this worker augmentation, direct task automation, or quality expansion? Explain.
- B. Estimate the productivity gains quantitatively where you can — for example, if System 3 saves 12-16 minutes per physician per patient encounter, and physicians see 20 patients per day, how much time is freed per physician per year? What could that time be used for?
- C. Does any of these systems represent a potential innovation spillover — a benefit beyond the hospital itself?
- PART 2 — LABOR-MARKET EFFECTS (Lessons 3-4)
- A. For each system, classify the tasks being automated as routine cognitive, routine manual, non-routine cognitive, or non-routine interpersonal. Does this classification predict how much displacement occurs?
- B. System 1 affects 340 administrative staff. Estimate how many positions might be eliminated, reduced, or transformed. Use the 70 percent automation estimate and supervisor ratio provided.
- C. What transition friction barriers would the displaced administrative workers face? Consider their age distribution (most administrative billing staff are 35-55), skills, and the local labor market.
- D. Are the new jobs created (AI supervisors, system maintenance, expanded clinical capacity) accessible to the displaced workers without significant retraining? Explain.
- PART 3 — INEQUALITY AND DISTRIBUTION (Lesson 5)
- A. Who captures the financial gains from these three systems — the hospital system, shareholders of the AI vendor, physicians, patients, or insurers? Trace each gain.
- B. Does the 12 percent improvement in early lung cancer detection disproportionately benefit any patient group? Consider which patients currently face underdiagnosis due to radiologist workload constraints.
- C. Does this deployment widen or narrow the skill premium within the hospital workforce?
- PART 4 — POLICY RESPONSE (Lesson 8)
- A. What obligations, if any, does the hospital have to the 340 at-risk administrative employees? Should it offer reskilling, severance, or priority rehiring for new roles?
- B. Should city or state government have any role in this transition — for example, if the hospital is the largest employer in the region?
- C. What governance structure would you recommend for ongoing AI quality assurance — who reviews the AI's diagnostic and billing decisions, and what recourse do patients have if an AI-assisted decision harms them?
- PART 5 — SYNTHESIS
- Write a 150-200 word recommendation memo to the hospital board. State clearly: (a) your overall recommendation (approve, approve with conditions, or reject), (b) the two strongest economic arguments in favor, (c) the two most serious economic risks, and (d) the single most important condition you would attach to approval.
- When your group is ready, present your Part 5 memo to the class. After each presentation, the class will identify one argument in favor and one risk the presenting group may have underweighted.
Real economic analysis rarely produces clean answers. A rigorous analysis names the assumptions you are making, identifies which conclusions depend on which assumptions, and tells decision-makers what additional information would change your recommendation. Confident assertions without acknowledged uncertainty are a sign of weak analysis, not strong expertise.
In the hospital scenario, the diagnostic AI reduces the radiologist headcount needed from 24 to 18. An economics student says 'this is pure displacement — 6 radiologist jobs are destroyed.' A second student says 'this is augmentation — the existing 18 radiologists can handle more volume and catch more cancers.' Which student is more correct?
A hospital administrator argues that the AI billing system will 'pay for itself' because it reduces administrative costs by $4.2 million per year. An economist reviewing the analysis says the administrator has omitted a significant cost. What is the most likely omission?