The impact of AI chatbots on EFL learners’ oral proficiency and willingness to communicate

The advancement of language education has promoted research into innovative teaching methods, including the integration of transformative technologies such as artificial intelligence (AI) chatbots. While AI chatbots have been extensively explored in second language (L2) learning, examining their impact on L2 oral proficiency and willingness to communicate (WTC) within the context of Informal Digital Learning of English (IDLE) remains essential for understanding how self-directed digital engagement can complement formal instruction and support L2 development in examination-oriented environments such as China (Lu et al., 2025). Therefore, this study investigated the effects of AI chatbot interaction on English as a Foreign Language (EFL) learners’ oral proficiency and WTC. Sixty EFL learners from a Chinese university were assigned to either a human–human interaction group or a human–AI chatbot interaction group. Oral proficiency was measured using elicited imitation tests (Wu & Ortega, 2013), while WTC was assessed through both self-reported questionnaires (Tai & Chen, 2023) and utterance frequency. After the four-week intervention, semi-structured interviews were administered to collect participants’ perceptions and attitudes regarding their use of the AI chatbot. Statistical analyses using one-way ANCOVA and two-way mixed-effects ANOVA demonstrated significant improvements in oral English proficiency and WTC in the human–AI chatbot interaction group. Interview findings indicated that learners valued the AI chatbot’s accessibility, authenticity, and multimodal features, which helped reduce anxiety and boost confidence. Regression analysis showed that WTC significantly predicted oral proficiency, highlighting the dynamic relationship between affective and behavioral factors in second language development.