English as a foreign language (EFL) educators have increasingly incorporated Artificial Intelligence (AI) into language learning; accordingly, this study explored the role of an AI-supported speaking tool in speaking skills, speaking anxiety, and willingness to communicate (WTC) among Chinese EFL students. Activity theory refers to a proper and applicable structure proposed for investigating how AI chatbots affect different dimensions. Applying this theory to the research allows for analyzing the complicated interactions between those factors and a comprehension of the effect of AI chatbots (EAP Talk) on learners’ WTC, speaking proficiency, and speaking anxiety. A total of 113 participants were initially recruited (58 EG, 55 CG); after excluding four multivariate outliers (all from the EG), the primary analytic sample comprised 109 learners (54 EG, 55 CG). Full-sample results ( N = 113) are also reported and were consistent in direction, though smaller in magnitude for speaking anxiety (see Sensitivity Analysis). Using pretests, the participants’ levels of speaking abilities, anxiety, and WTC were assessed prior to the intervention. Subsequently, the EG received AI-supported speaking practice through EAP Talk, an AI-based speaking assessment and feedback application that facilitated iterative practice and immediate automated feedback, whereas the CG was taught the same material without the use of this tool. Using post-tests following the treatment, both groups were evaluated on their speaking skills, anxiety, and WTC. The results revealed notable posttest differences between the two groups: WTC showed a medium effect size, speaking anxiety showed a large effect size, and speaking skill showed a very large effect size. Overall, the results suggest that the EAP Talk intervention was associated with improved speaking performance and WTC, as well as reduced speaking anxiety among the participating EFL students. Interpreting the findings through the lens of Activity Theory helps situate the results within the broader relationships among learners, tools, and instructional context; however, the theory is used here as an interpretive lens rather than as a directly tested explanatory model.
No takes yet. Share an insight, caveat, or question.
Xue Wang (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: