AI Ethics Case Court

Overview

Schools and colleges are already making consequential decisions with AI: scoring writing, flagging suspected cheating, monitoring student devices, predicting who might drop out. These uses can save time and catch real problems, and they can also harm the students with the least power to push back. In this mock-court session, small courts argue fictional cases using an ethical principles bench card, then rule "allowed," "allowed with conditions," or "not allowed." The conditions they write become a draft of practical guidance for their own campus.

Objectives

• Participants will identify ethical issues (privacy, fairness, transparency, accountability, human oversight, learning benefit) in realistic AI uses in education.
• Participants will argue both sides of an AI use case using evidence from the scenario rather than general opinion.
• Participants will propose specific, enforceable conditions that would make an AI use acceptable, or explain why no conditions would.

Step by step

1. 10–5 min

   Hook

   Would you sign it?

   Read aloud: "Starting Monday, an AI system will flag any student essay it judges to be AI-written. Flagged essays receive a zero pending review." Ask: "Stand up if you'd sign this policy as written." Most stay seated. Ask one person why. Then: "Today you'll figure out what it would take, if anything, to make a policy like this defensible."

   Facilitator noteSomeone usually mentions false accusations. Acknowledge it and hold it for Case 2.

2. 25–13 min

   Model

   The bench card

   Walk through Handout B's seven principles with one quick example each. Emphasize two questions judges must always ask: "Who is accountable when it's wrong?" and "Can the person affected see, question, and appeal the decision?" Explain the three verdicts: Allowed, Allowed with conditions (conditions must be specific enough to check), or Not allowed.

   Facilitator noteStress that "ethical" doesn't mean "never." Many strong rulings are "allowed with conditions." Many weak ones are vague ("use responsibly").

3. 313–25 min

   Explore

   Prepare the case

   Courts receive a case file. Prosecution builds the strongest case that this use harms students or violates a principle. Defense builds the strongest case for the benefits and argues any harms can be managed. The bench studies the bench card and writes three questions they'll ask each side. Everyone must cite details from the case file, not general feelings about AI.

   Facilitator noteAssign sides randomly. Arguing the side you disagree with is where the learning happens, especially for leaders who will face both arguments from their community.

4. 425–43 min

   Practice

   Court is in session

   Run the trial: prosecution opens (2 min), defense opens (2 min), bench questions both sides (6 min), each side gives a 1-minute closing. The bench deliberates aloud for 4 minutes while others listen, using Handout B, and completes Handout C: the verdict, the principle that weighed most, and the conditions. Bang the gavel.

   Facilitator noteLook for benches that separate the tool's reliability from the decision rule built around it. A flawed tool used only to start a conversation is very different from the same tool used to assign a zero.

5. 543–55 min

   Debrief

   Compare rulings

   Each court reads its verdict and top two conditions aloud. Post conditions on "Conditions we'd adopt." Ask the room: "Which conditions showed up across different cases?" Common ones: a human makes the final decision, students and families are told, there's a real appeal process, data is minimized and deleted, the tool is checked for unequal error rates. Then: "Who might be left out or harmed by our own conditions?"

   Facilitator noteOn detectors, make sure the room hears the accurate note: AI-writing detectors are known to be unreliable, and there are published concerns about false positives for non-native English writers. Don't cite a specific error rate; none is universal.

6. 655–70 min

   Transfer

   From verdict to guidance

   Mixed groups (one person from each court) draft three sentences of guidance for their campus, department, or course using the posted conditions: one "We will…", one "We will not…", and one "If AI contributes to a decision about a student, the student can…". Leaders take these to their next policy conversation; faculty adapt them for a syllabus or department handbook.

   Facilitator noteLeaders can run the five-question check on the draft: vision, evidence, change in instruction, who might be left out, and how families will understand.

Adaptations

Higher Ed faculty

Use Cases 3 (remote proctoring) and 6 (faculty use of AI in recommendation letters) and add accreditation and disability services considerations to the bench card.

District and campus leaders

Add a "procurement" witness who must answer what the vendor contract says about data use, retention, and deletion. Leaders rule on the contract, not just the practice.

Grades 9–12 students

Teachers can run Cases 2 and 4 with students, who argue about decisions that directly affect them. Their conditions make excellent input for a student AI-use agreement.

Pairs well with

• Generative AI Adoption ChecklistTurn your court's conditions into a campus or district adoption review.
• PROTECT Rubric v2.0Useful for rulings that hinge on privacy and data protection.