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

License

Grant ID

Call number

Persistent link to this record