Adoption and Integration of Artificial Intelligence in Foreign Language Teaching: Multi-Institutional Analysis of Higher Education

This study examines the adoption and integration of artificial intelligence (AI) tools in foreign language teaching (FLT) in 13 higher education institutions. A total of 327 faculty members and instructors from the School of Languages participated in a quantitative, cross-sectional survey. A purpose-designed Likert-type instrument was employed to measure five key dimensions: Instructional Integration, Material Development, Assessment and Feedback, Self-Efficacy, and Perceived Barriers. In this study, AI adoption was operationally defined as the self-reported behavioral Integration of AI tools into pedagogical practice, distinguishing it from related constructs such as attitudes, self-efficacy beliefs, and perceived contextual barriers, which were treated as predictors and correlates. The results indicated a lower-mid level of self-reported AI adoption (overall M = 2.41, SD = 0.89), highlighting a significant discrepancy between participants’ awareness of AI tools and their pedagogical implementation. The highest adoption was observed in Material Development (M = 2.80), whereas Assessment and Feedback yielded the lowest scores (M = 2.10). Multiple regression analysis identified a lack of institutional training (β = –.31, p < .001), technostress (β = –.24, p < .01), and perceived threat to professional autonomy (β = –.18, p < .01) as the strongest predictors of low integration. These findings suggest that, despite widespread access to AI technologies, substantive pedagogical Integration in FLT remains constrained by infrastructural, psychological, and epistemological barriers. The implications for institutional policy, professional development, and curriculum design in language education are discussed.