Reshaping the Language Learning Landscape: Factors Influencing Learners' Use of Open‐Source Text‐to‐Video AI (Open‐Sora) in Second Language Acquisition

ABSTRACT Background Following the disruptive impact of ChatGPT on language learning, the emergence of Sora has once again attracted considerable attention in the field. This novel generative AI product offers new learning modes and resources, bringing new possibilities for the transformation of language acquisition. However, research systematically investigating learners' intention to use such text‐to‐video AI tools remains scarce. Objectives This study utilises text‐to‐video AI‐generated teaching materials produced using Open‐Sora (an open‐source implementation inspired by Sora, not the unreleased official Sora model) as experimental content, aiming to explore college students' intention to use this system for second language (L2) learning and their perceptions of the technology. Methods In response, this study, grounded in the stimulus–organism–response framework, employs the Information System Success Model, the Task‐Technology Fit model, and the Unified Theory of Acceptance and Use of Technology. Data were collected from 335 university students at universities in Western China and analysed using a structural equation model. Results and Conclusions Performance expectancy, L2 enjoyment (LEE), and self‐efficacy (SE) significantly and positively predict Behavioural Intention (BI) to use Open‐Sora, whereas effort expectancy and facilitating conditions do not significantly predict BI. Furthermore, SE partially mediates the relationship between LEE and BI. This study enhances our understanding of college students' interactions with text‐to‐video AI in L2 learning contexts and provides implications for the educational use of such AI technologies.