Innovations in Science and Technologies Volume 3 Issue 8 (2026) · pp. 95-103
A METHODOLOGICAL FRAMEWORK FOR ENHANCING EFL LEARNERS’ SPEAKING AND WRITING COMPETENCE THROUGH ARTIFICIAL INTELLIGENCE TECHNOLOGIES
Kholboeva, Durdona, Khamzaev, Jamshid
Abstract
Developing speaking and writing competence remains one of the most persistent challenges in English as a Foreign Language (EFL) instruction, since traditional classrooms rarely provide the individualized practice and immediate feedback these productive skills require. Building on the authors' earlier conference thesis, which surveyed AI-based tools such as chatbots, natural language processing (NLP) applications, and speech-recognition systems, this article develops that work into a structured methodological framework intended for classroom implementation. Using a qualitative literature-synthesis design, the study analyzed sixteen empirical and meta-analytic sources published between 2021 and 2026 to determine which categories of artificial intelligence tools are most consistently associated with measurable gains in speaking fluency, pronunciation accuracy, writing accuracy, writing organization, and learner motivation. The analysis consolidated the instructional sequences reported across these studies into a five-stage cycle covering AI-mediated input, interactive production, automated feedback, learner self-revision, and adaptive re-practice, designed to structure classroom use of AI tools rather than leaving such use incidental. The findings indicate that chatbot- and speech-recognition-based tools are most strongly associated with oral fluency and pronunciation gains, while natural-language-processing checkers and generative AI tools are most strongly associated with writing accuracy and organizational quality; generative AI shows the broadest reported influence on learner motivation and autonomy. The discussion situates these findings against the reviewed literature, notes implementation constraints such as unequal technological access and learner over-reliance on automated corrections, and proposes recommendations for lesson design, teacher preparation, and future empirical validation of the proposed cycle in university EFL classrooms in Uzbekistan.
artificial intelligenceEFL methodologyspeaking skillswriting skillschatbotsnatural language processingspeech recognitiongenerative AIadaptive learninginstructional design.
Metadata source: the journal's OAI-PMH archive · Sindex does not store the full text; it links to the source.