.png){kind=icon}
.png){kind=icon}
Return
Exploring L2 Learner Engagement in Spoken English Practices with a GenAI-powered Chatbot
.png){kind=icon}
PRE
AI
S
Y
J
DOI：10.1016/j.system.2026.103982.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Generative artificial intelligence (GenAI) engages students in targeted and context-aware interactions, holding the potential to revolutionize chatbot-supported second language (L2) learning. However, GenAI-assisted language learning research has been inadequate in addressing learner engagement that depicts the entire learning process. This study explored 89 Chinese English learners’ engagement when they practiced spoken English with E-Talk, a GenAI-powered chatbot, under different pedagogical designs. Data sources included engagement questionnaire, human-GenAI chat logs, and semi-structured interviews. Analyses revealed that the teacher guidance group demonstrated significantly higher behavioral, cognitive, and social engagement compared to the regular group, revealing the value of structured instructional support. Notably, cognitive engagement was jointly shaped by teacher guidance and task mode, as thematic dialogue elicited higher cognitive engagement in the teacher guidance group. Taken together, the results highlighted the importance of pedagogical design and the context dependency of learner engagement in GenAI-assisted spoken English practice. This study provided empirical evidence supporting the role of teachers and task design in enhancing GenAI-powered L2 learning experience and contributed to a wider discussion on the importance of pedagogical design to facilitate GenAI’s role in individualized learning.
Journal
IF:
5.6
Papers:
482
Citations:
9.7K