Development and Validation of an AI-Assisted E-Comic to Support Deep Learning and Critical Thinking in Senior High School
Main Article Content
Abstract
Indonesian students consistently perform below the OECD average in reading and scientific literacy, a gap that underscores the need for instructional media that make abstract, conceptually dense topics accessible while cultivating higher-order thinking. This study developed and evaluated an artificial intelligence (AI)-assisted electronic comic (e-comic) designed to operationalise the deep learning paradigm (mindful, meaningful, and joyful learning) for senior high school students. Using the ADDIE (Analysis, Design, Development, Implementation, Evaluation) research and development model, the e-comic was produced through generative AI tools (a large-language-model chatbot for narrative and prompt design, and AI-assisted illustration platforms for character-consistent artwork) around three narratives addressing biology and language topics that teachers had identified as difficult to teach. Content and media validity were assessed by expert validators using a four-point Likert rubric; practicality and reliability were assessed through a 17-item student questionnaire (Cronbach's α = 0.82); and learning impact was assessed with a pretest–posttest design (N = 30). All content-validity indicators reached the maximum score (4/4, 'very valid'), and media validity averaged 3.67/4. The questionnaire showed a modal item score of 3 ('valid/practical') on 16 of 17 items, with an overall mean of 3.16 (SD = 0.53) on a 4-point scale. Students' comprehension and critical-thinking scores rose from a pretest mean of 46.40 (SD = 12.61) to a posttest mean of 82.10 (SD = 8.54), a statistically significant gain, t(29) = 17.69, p < .001, Cohen's d = 3.31, corresponding to a moderate-to-high normalised gain (⟨g⟩ = 0.67). Qualitative feedback indicated that students valued the clarity of illustrations and the accessibility of the narratives, while some requested additional chapters, colour illustration, and reduced reliance on AI-generated art style. The findings suggest that AI-assisted e-comics are a valid, reliable, and practical medium for strengthening critical thinking within a deep learning framework, while highlighting specific directions, visual diversification and curriculum alignment for iterative refinement.


