The Use of AI-Generated Comic Strips Based on Indonesian Local Culture to Enhance Students' Vocabulary Achievement: A Study at the Eighth Grade Students of SMP Negeri 4 Semarang

Abstract

Vocabulary mastery remains a major obstacle for Indonesian junior high school EFL learners, who often struggle to understand and use English words in meaningful contexts because instruction still relies heavily on memorization and translation. This study investigated whether AI-generated comic strips based on Indonesian local culture (Kecak Dance, Tenun Ikat, and Wayang Kulit) could enhance eighth-grade students' vocabulary achievement. A pre-experimental one-group pretest-posttest design was employed involving 32 students of class VIII A at SMP Negeri 4 Semarang in the 2025/2026 academic year, using a 25-item multiple-choice vocabulary test (Cronbach's Alpha = 0.880). The mean score increased from 45.00 in the pretest to 75.75 in the posttest. Because the pretest data were not normally distributed (Shapiro-Wilk, Sig. = 0.000), a Wilcoxon Signed Rank Test was applied, yielding an Asymp. Sig. (2-tailed) of 0.000 (p < 0.05), with all 32 students showing improvement and none regressing. These findings indicate that AI-generated comic strips grounded in familiar local-cultural themes significantly improved students' vocabulary achievement by combining visual storytelling, contextual exposure, and culturally relevant content, supporting Dual Coding Theory and Multimedia Learning Theory. The study implies that culturally responsive, technology-assisted visual media can serve as an effective alternative medium for vocabulary instruction in EFL classrooms.