Generative AI-based digital storytelling enhances cross-cultural symbol learning through learning engagement

Abstract

Introduction Cross-cultural learning of traditional cultural symbols is pedagogically demanding because learners must move beyond visual recognition toward historically situated, affectively meaningful, and transferable interpretation. This study examined whether generative artificial intelligence (GenAI)-based digital storytelling improves cultural-symbol learning through learning engagement. Methods Study 1 randomized 276 university students to expository text, conventional digital storytelling, or GenAI-based adaptive digital storytelling. Study 2 surveyed 487 students who had recently completed AI-supported cultural learning activities and used confirmatory factor analysis and structural equation modeling to test the proposed engagement pathway. Results In Study 1, the GenAI condition produced higher learning engagement, F(2, 273) = 24.37, p < 0.001, ηp2=0.151, and higher cross-cultural learning outcomes, F(2, 273) = 18.92, p < 0.001, ηp2=0.122. Engagement partially mediated the GenAI effect (B = 0.32, 95% CI [0.19, 0.47]). In Study 2, GenAI storytelling quality predicted engagement (β = 0.61, p < 0.001), engagement predicted learning outcomes (β = 0.47, p < 0.001), and the indirect effect was significant (β = 0.29, 95% CI [0.22, 0.37]). Discussion Across complementary designs, the findings indicate that GenAI-based digital storytelling can support cross-cultural symbol learning by creating more engaging adaptive narrative experiences. Study 2 provides convergent mechanism evidence rather than independent causal proof.