Artificial intelligence-assisted speaking in Spanish as a foreign language: effects on foreign language emotions and willingness to communicate

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Introduction Although AI-powered conversational tools are increasingly used in foreign language education, little is known about their longitudinal effects on learners’ emotional experiences and willingness to communicate (WTC), particularly among beginner learners of languages other than English. Grounded in Positive Psychology, Control-Value Theory, and the WTC framework, this study examined the effects of AI-assisted speaking practice using TalkPal AI on foreign language enjoyment (FLE), foreign language anxiety (FLA), foreign language boredom (FLB), and WTC in Spanish as a foreign language. Methods This longitudinal quasi-experimental study involved 69 Chinese university students with no formal prior learning experience in Spanish, who were assigned to an experimental group ( n = 35) and a control group ( n = 34) based on intact classes. Quantitative data were collected at three time points using validated questionnaires. Semi-structured interviews with nine purposively selected students from the experimental group were analyzed thematically to contextualize the quantitative findings. Quantitative data were analyzed using linear mixed-effects models, planned contrasts, and change-score correlation analyses. Results No significant short-term group differences were observed during the early stage of the intervention. Across the full intervention period, the experimental group showed a significantly greater reduction in FLB than the control group. At the final measurement point, the experimental group reported significantly higher WTC and significantly lower FLB, while FLE showed a positive but non-significant tendency and no significant group difference was observed for FLA. In the experimental group, increases in FLE were positively associated with increases in WTC over time, whereas changes in FLB and FLA were not significantly related to WTC. Qualitative findings further showed that learners gradually reported greater enjoyment and engagement and less monotony as AI-mediated interaction became routinized over time. Learners also perceived AI-supported speaking practice as a psychologically safe rehearsal space, although evaluation-related anxiety and the limited emotional authenticity of AI interaction remained salient. Discussion These findings suggest that AI-assisted conversational practice may exert delayed and differentiated effects on beginner learners’ emotional experiences and willingness to communicate, primarily by reducing boredom and supporting WTC rather than alleviating anxiety.