An umbrella review on conversational AI in language learning (CoALL)

• In Event: AERA Roundtable Session Saturday 7:45 am - Gold 3
  In Roundtable Session: AI supported writing: Tracing engagement and learning

Abstract

This umbrella review synthesizes the fragmented research on conversational AI in language learning (CoALL) by analyzing 37 reviews that cover primary studies from 1999 to 2024. Guided by social cognitive theory, we discussed the affordances of CoALL, its impact on different educational levels, and the collaboration between teachers and conversational AI. Findings reveal CoALL’s capacity to facilitate authentic dialogue, provide adaptive formative feedback, and support the psychological needs of language learners in K-12 and higher education. All levels of language education benefit from CoALL, with younger learners benefiting more, although these differences were not significant. Evidence of teacher-AI collaboration is limited but growing, raising questions about the adaptations required and the roles of teachers and AI.

Authors

• Weijiao Huang, University of Hong Kong

• Xinyuan Yang, University of Hong Kong

• Luke K. Fryer, University of Hong Kong