Key points are not available for this paper at this time.
The widespread adoption of artificial intelligence (AI) tools in English language education has created new opportunities and challenges for students in global contexts. This mixed-methods study investigates Kosovan English as a Foreign Language (EFL) students’ acceptance of AI tools for English learning, examining their practices, perceptions, challenges, and suggestions for improvement. Using the Technology Acceptance Model framework, a comprehensive questionnaire was distributed to 85 university-level EFL students in Kosovo, representing an understudied context in AI-education research. The study explores students’ usage patterns of popular AI tools alongside their multidimensional attitudes toward effectiveness, usability, trust, and behavioral intentions. Quantitative analysis of Likert-scale items was complemented by qualitative thematic analysis of open-ended responses regarding benefits, challenges, and desired improvements. Kosovan EFL students find AI tools easy to use and valuable for vocabulary, grammar, and writing, yet significant trust concerns about AI feedback reliability persist. Students frequently use AI tools for vocabulary learning, translation, and writing assistance. Challenges include accuracy concerns, overdependence, and limited contextual understanding; improvements focus on reliability, cultural adaptation, and pronunciation features. The study contributes to understanding AI acceptance in understudied educational contexts in Kosovo and highlights the complex relationship between usability and trust in educational technology adoption. This study innovates by empirically documenting the trust – usability paradox through visualization, focusing on an understudied post-conflict context, and linking specific usage practices to student perceptions. These findings suggest that developers and educators should prioritize transparency, cultural relevance, and balanced integration of AI tools in language curricula.
Memidin Braha (Tue,) studied this question.