2025
Listening with Language Models: Using LLMs to Collect and Interpret Classroom Feedback
LLMs · Conversational AI · Learning analytics
This project explores the use of LLM-powered chatbots as an alternative to traditional end-of-quarter surveys, engaging students in reflective, conversational dialogues mid-course. The result is a three-part system: PromptDesigner, which allows instructors to craft and test course-specific feedback prompts; FeedbackCollector, a student-facing GPT-4 chatbot that elicits detailed, contextual reflections; and FeedbackAnalyzer, a dashboard that synthesizes student responses into actionable insights for instructors. A pilot study across two graduate courses at UC Santa Cruz found that the system surfaced richer, more specific feedback than standard surveys, enabled real-time course adjustments, and was more engaging for students. The project is also investigating the system’s potential for tracking soft skills, including communication, collaboration, goal setting, and conflict resolution, in group work contexts.