| Course Subject: | Law 7410, Domestic Violence Clinic Law 6103, Criminal Justice |
| Student Level: | Graduate |
| Number of Students: | 24 (total across two courses) |
| Developed by: | Hemanth C. Gundavaram, Associate Dean for Academic and Faculty Affairs, School of Law |
What Students Will Do
Students were given a case study to analyze and questions to answer. They posted their responses to the questions in a written discussion in Canvas. Their responses were anonymized and uploaded to AI to generate a podcast that synthesized their collective analyses.
Purpose
The overarching purpose for initiating this practice was to explore AI’s versatility across different legal education contexts—from doctrinal law school courses to law school clinical settings. It was piloted in two courses, a first-year online Criminal Law course and an upper-level on-ground Domestic Violence Clinic. The goal of this activity was to experiment with using AI to synthesize crowdsourced student knowledge and share the output back in audio format with students, providing them with a new vantage point and modality for meaning-making in case study discussions.
Note: This AI practice was developed in collaboration with Professors Daniel Medwed in his Criminal Justice course and Hayat Bearat in her Domestic Violence Clinic.
Assessment
This was a pilot project, and therefore assessment focused on faculty and student perceptions of the experience. Student and faculty feedback was gathered and described below.
Student and Faculty Reflections
The two implementations of this practice yielded starkly different outcomes based on the course context and subject matter of the case. Criminal Law students had a positive experience, reporting that they gained a “deepened understanding of the necessity defense,” which is a doctrinal concept in the area of criminal law. Students requested expansion to other case study discussions. In contrast, upper-level Domestic Violence Clinic students found that the podcast format, with significant banter, “seemed to make light of a topic that was quite serious,” which was related to media portrayals of domestic violence survivors. They also felt the synthesis failed to capture their extensive engagement with the material. The professor concluded that “it did not seem to aid in developing critical thinking” and that in-person live discussion was better suited for this work
As the faculty member who coordinated the experience across the two courses, I concluded that these parallel experiments provided crucial insights about AI’s appropriate use in legal education. The success differential between doctrinal and clinical contexts was striking—AI effectively synthesized legal analysis but failed with trauma-informed content. I recognized that some educational practices—particularly those requiring vulnerability and emotional processing — cannot be effectively automated. It’s important to consider the context and course context when implementing innovations. The contrasting outcomes demonstrate that AI’s promise must be balanced with clear understanding of its limitations.
Step-by-Step Directions
| Step 1 | Students read a case study scenario, related to a specific concept of law, and received a discussion prompt to guide their analysis. (see prompts in related materials below) Students then wrote a response to the prompt. |
| Step 2 | Written responses to the prompt were downloaded, anonymized and aggregated. |
| Step 3 | With the permission of students, anonymized responses were uploaded into Google LM with a corresponding prompt (see prompt in related materials). |
| Step 4 | Google LM created an audio podcast based on a synthesis of the responses, which was shared with the students to inform additional discussion of the case. |
| Iteration | Student feedback was gathered on their perceptions of the podcast. Students in the first-year course found it deepened understanding. Students in the upper-level course found that the informality and tone of the podcast was inappropriate to the content of the case and was, thus, distracting. |
Related Materials
Download PDF version of this assignment
Return to AI Gallery Home