Using AI to Design Public-Service Cases That Require Student Judgment

• In Event: AI for Teaching, Learning, and Student Research: Helping educators and students use AI while preserving independent judgment

Abstract

Public administration instructors need cases that connect with students’ experience and require them to apply concepts, weigh competing public values, and defend consequential choices. This interactive lab demonstrates how instructors can use generative AI to develop such cases while retaining responsibility for disciplinary accuracy and assessment design. Drawing on an undergraduate leadership assignment at Tarleton State University, the workflow begins with learning objectives and the reasoning students must demonstrate. An instructor-reviewed, anonymous summary of students’ backgrounds and interests—including summaries informed by Canvas’s AI discussion-summary feature where available—helps identify a relevant setting. Participants will see how instructors can guide and revise AI-generated scenarios to establish realistic constraints, plausible alternatives, and meaningful tradeoffs. The example asks students to choose how a county agency should implement a new service system under a fixed deadline. Informed by team-based learning, the live activity asks students to make an individual provisional choice, deliberate in small teams on the same problem, commit to a collective decision, and compare reasoning across teams. Students apply leadership concepts, examine the strongest alternative, identify an unresolved consequence, and reflect on their leadership tendencies. An asynchronous “Path B” assesses these core analytical skills through an individual decision video, followed by a debrief presenting competing arguments. Common rubric criteria connect the pathways while recognizing the distinct learning opportunities offered by team deliberation. The session distinguishes instructor AI use for assignment design from permitted student assistance with brainstorming and organizing their own analysis. Participants will consider how instructions, disclosure requirements, and rubric criteria support accountable AI use without treating a polished response as proof of independent reasoning. The 20-minute lab includes three minutes of framing, seven minutes of demonstration, six minutes of participant revision and peer comparison, and four minutes of discussion. Participants will revise a weak AI-generated case option and explain how their revision strengthens the judgment students must exercise. Participants will practice specifying pedagogical constraints in an AI prompt, evaluating the resulting alternatives, and connecting assessment requirements to learning objectives. They will receive an adaptable prompt sequence, assessment example, and review checklist for use in their own courses. No coding or prior AI expertise is required.

Author

• Marcela Nava, Tarleton State University
  Facilitator