Perceptions and Acceptance of Generative Artificial Intelligence Influencing Chinese EFL Learners’ Engagement in Informal Digital Learning of English: Mediating Roles of Self-Efficacy and Motivation

Generative artificial intelligence (GenAI) has emerged as a powerful tool in the Informal Digital Learning of English (IDLE) environment, offering personalized, interactive, and innovative language learning experiences that may enhance learning engagement. However, the influencing path of perceptions and acceptance of GenAI on the learning engagement of English-as-a-foreign-language (EFL) students has not been sufficiently explored. Specifically, little is known about how cognitive and motivational factors jointly mediate this relationship. Self-efficacy in language learning has been found to play a critical role in regulating motivation through resilience, perseverance, and goal determination, thereby influencing the quality and quantity of effort that students invest in engaging with GenAI-assisted IDLE. This study investigated the effects of GenAI perceptions and acceptance on IDLE engagement among 369 university EFL learners. Using structural equation modeling (SEM), the study analyzed both the direct effects of perceptions and acceptance of GenAI on learning engagement and the serial mediating roles of self-efficacy and motivation. The results showed that (1) GenAI perceptions and acceptance exerted a significant and positive direct effect on learners’ engagement; (2) motivation significantly mediated the relationship between GenAI perceptions and acceptance and engagement; and (3) self-efficacy indirectly enhanced engagement by increasing motivation, forming a chain mediation pathway. These findings provide insights into how self-efficacy and motivation jointly mediate the relationship between GenAI perceptions and acceptance and learning engagement and offer guidance for the implementation of GenAI-enhanced tools and personalized learning strategies in fostering student autonomy and optimizing engagement in IDLE environments.