Design, Development and Evaluation of a Two-Level AI-Generated Quiz Game for Teaching Multimedia Elements using Gemini Canvas

Abstract

Foundation-level multimedia courses require students to master the five core media elements, text, graphics, audio, video, and animation, yet abstract terminology and formats are often difficult to retain through lectures alone. This study reports the design, development, and evaluation of a two-level, AI generated quiz game built with Gemini Canvas to support teaching multimedia elements in the Multimedia System Foundation course. The game was developed using the ADDIE model and organised into five element modules, each offering an easy recall level and a hard application level, with instant explanatory feedback after every item. We used a one-group pre-test/post-test design with 80 undergraduate students, complemented by a perception survey grounded in the Technology Acceptance Model. Paired-samples analysis showed a significant improvement from pre-test to posttest, rising from a mean of 10.29 to 15.51 out of 20, t(79) = 40.35, p < .001, with a normalised gain of 0.54, indicating a medium learning effect. Gains were larger for application than for recall. Learner perceptions were highly positive across perceived usefulness, perceived ease of use, behavioural intention, and engagement, with all construct means ranging from 4.29 to 4.37 on a five-point scale. The findings indicate that an AI-generated quiz game is a well-accepted supplement associated with improved post-test performance in teaching multimedia elements, and they demonstrate a practical, low-cost workflow for educators to produce classroom games using generative AI tools.