Students' use of an LLM-Powered virtual teaching assistant for recommending educational applications of games
Date
2025
Authors
Shah, M.
Pankiewicz, M.
Baker, R.S.
Chi, J.
Xin, Y.
Shah, H.
Fonseca, D.
Editors
Plass, J.L.
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Conference paper
Citation
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2025 / Plass, J.L. (ed./s), vol.15259 LNCS, pp.19-24
Statement of Responsibility
Conference Name
10th Joint International Conference on Serious Games (7 Nov 2024 - 8 Nov 2024 : New york city, Ny)
Abstract
In this paper, we discuss the application of JeepyTA- a virtual teaching assistant (TA) powered by a large language model (LLM) - within a graduate-level course on games and learning in Spring 2024. Specifically, we describe how Jeepy-TA was integrated in two "play journal" assignments where twenty-two students documented their observations of playing classic and contemporary digital games, shared insights from assigned readings, and proposed educational applications of the respective games through consultation with JeepyTA. We (a) report students' overall communication patterns with JeepyTA and (b) identify five ways in which students incorporated JeepyTA's recommendations for Minecraft. We conclude this paper with directions for future research and practice on supporting students in using LLMs for reflecting on their experiences and creatively adapting games for learning.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
Copyright 2025 Springer