AI-assisted L2 learning attitudes and willingness to communicate: the mediating role of language anxiety

Abstract

This study investigated the mediating role of skill-based foreign language anxiety (FLA) in the relationship between attitudes toward AI-assisted second language (L2) learning and willingness to communicate (WTC) among Egyptian English as a foreign language (EFL) student teachers. Using a cross-sectional correlational design, 853 participants from Al-Azhar University completed the AI-assisted L2 learning attitude scale (AL2AS), the skill-based foreign language learning anxiety scale, and the WTC scale. Mediation analysis using the PROCESS macro with 5,000 bootstrap samples revealed that AI-assisted learning attitudes positively predicted WTC (β=.288) and negatively predicted skill-based FLA (β=-.206), while skill-based FLA negatively predicted WTC (β=-.202). The indirect effect through skill-based FLA was significant (β=.042, 95% confidence interval (CI) [.020, .068]), indicating partial mediation accounting for 14.3% of the total effect. This study focuses on student teachers in the Egyptian EFL context, distinguishing anxiety patterns across language competencies using a skill-based anxiety measure. It reveals that positive attitudes towards artificial intelligence (AI) enhance communicative readiness by reducing anxiety. Although the cross-sectional design limits causal inference, the findings advocate integrating AI tools into teacher education to create supportive environments that promote communicative competence, addressing a gap in research in this area.