Generative AI in Digital Storytelling: An Experimental Study of Junior High School Students’ Digital Story Quality, Creative Process Patterns, and Creative Thinking

Description

Digital storytelling (DST) provides an authentic context for developing creative thinking and multimodal expression, yet evidence on generative artificial intelligence (GAI) in K–12 settings remains limited, particularly at the process level. This pretest–posttest randomized controlled study involved 94 seventh-grade students assigned to an AI-supported group (AIG) or a non-AI group (NAIG). The AIG used GAI for planning, text and material generation, and revision, whereas the NAIG used non-AI digital tools. Data included digital stories, creative thinking tests, and screen recordings. ANCOVAs, Mann–Whitney U tests, and lag sequential analysis were conducted. The AIG showed significantly higher overall digital story quality (adjusted M = 22.27 vs. 17.79, partial η² = .195), with advantages in structure, accuracy, completeness, appearance, and innovation, but not interaction. The NAIG showed higher frequencies of writing and multimedia integration, whereas the AIG showed more frequent revising sources. Lag sequential analysis revealed distinctive transitions linking planning, writing, evaluation, and revision in the AIG. The AIG also showed significantly higher overall creative thinking (adjusted M = 9.12 vs. 8.02, partial η² = .110), particularly in text creativity and creative problem solving, but not idea generation. These findings highlight differences in both creative outcomes and process organization under GAI-supported DST.

Institutions

• Shaanxi Normal University

  Shaanxi, Xi'an

Categories

Licence