- Stoyka, Olesia (orcid.org/0000-0002-7695-6100) and Dupyn, Yelyzaveta (orcid.org/0009-0005-0061-2490) (2026) The impact of ai on language learning: between real progress and the illusion of fluency Continuing professional education: theory and practice, 3 (88). pp. 162-173. ISSN 2412-0774
|
Text
Стойка, Дупин.pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (440kB) |
Abstract
The rapid integration of generative artificial intelligence (AI) into English as a Foreign Language (EFL) education has prompted growing debate over whether AI enhances genuine language competence or merely creates an illusion of fluency. Recent studies and the OECD have emphasized both the pedagogical potential of AI for personalized learning and adaptive feedback and the methodological risks associated with learner overreliance on AI-generated output. The aim of this study is to determine how AI influences English language learning in comparison with traditional instruction and to identify the conditions under which AI supports authentic language development. The objectives include reviewing recent research on AI-supported EFL learning, comparing AI-assisted and conventional instruction, operationalizing the concept of the illusion of fluency. The study employed a comparative pedagogical design combining theoretical analysis, a quasi-experimental approach, pre- and post-testing, learner questionnaires, descriptive statistical analysis, and an independence-of-performance criterion. Forty A2–B1 undergraduate learners (aged 17–22) participated in a four-task instructional cycle. The experimental group used ChatGPT and similar chatbot-based tools for preparation, revision, and feedback, whereas the control group followed teacher-led instruction without AI support. The findings indicate that AI enhances selected aspects of language learning but may simultaneously inflate learners’ perceptions of their own competence. Consequently, AI should be integrated within a staged blended-learning model in which it supplements, rather than replaces, independent language production and teacher guidance. Further research should investigate the illusion of fluency through larger longitudinal and cross-cultural studies, examine AI literacy as a moderating factor, refine AI-informed assessment practices, and explore AI-supported development of pronunciation and listening skills.
Downloads
Downloads per month over past year
Actions (login required)
![]() |
View Item |


