exploring the effects of chatbot-assisted language learning on learners’ language achievement and sense of metavolvement

Artificial intelligence (ai) is changing language education through personalized and interactive practice. however, its impact on language achievement and emotional aspects of learning is unclear. therefore, this mixed-methods study investigated the effects of chatbot-assisted language learning (chall) on learners’ english language achievement and sense of metavolvement—the deepest level of metacognitive engagement. in phase 1, the sense of metavolvement scale (soms) was developed and validated. in phase 2, following a quasi-experimental design, 44 intermediate iranian efl learners were divided into an experimental group (n=22, using the ai chatbot pi) and a control group (n=22, receiving conventional classroom instruction). both groups had pre- and post-tests for language achievement and metavolvement, and the experimental group sat for semi-structured interviews post-intervention. there were significant gains in the experimental group’s in-class metavolvement and language achievement. there were also significant differences between the two groups in post-test metavolvement, while there was no significant difference in their post-test language achievement. furthermore, interview data illustrated chall as a low-anxiety practice context with cultural and pragmatic limitations. therefore, chall can complement but not replace conventional instruction.