Multimodal AI Tools as Mediators of Linguistic Development in EFL Academic Writing

Read full text
Authors: Oleksandr Tomchakovskyi, Yuliia TOMCHAKOVSKA, Lesia Strochenko, Lidiia Shvelidze, Svitlana Yeryomenko
Cited by: 1

Abstract

Despite the growing adoption of artificial intelligence tools for drafting, revising, and editing of academic texts, little empirical evidence is available with respect to how multimodal AI-tools mediate linguistic development beyond being merely corrective or generative mechanisms. The research focuses on exploring the impact of current state-of-the-art multimodal AI systems on the development of academic writing proficiency among English as a foreign language learners at the National University “Odesa Law Academy,” particularly in terms of linguistic development, genre knowledge, and agency. The research provides evidence on the effects of multimodal AI on critical aspects of academic writing, including the following: the sophistication of word-level variabilities, syntactical correctness and complexity; and the development of cohesion, coherence, and genre compliance and meta-linguistic comments during and after the writing process in academic texts. The research uses a mixed research methodology to compare student academic texts and evaluation on an additional perspective on patterns of interactions emerging from text, audio, and graphic research techniques and experimentation. It gives evidence that pedagogically mediated multimodal AI use is associated with enhanced academic vocabulary range, improved syntactic organization, and more coherent, genre-consistent writing. Interaction analysis shows how linguistic development is enabled via cross-modal transformations, visual feedback, iterative prompting, and oral-written mediation that together foster noticing and scaffolded revision. At the same time, the study highlights several risks related to over-reliance on AI, homogenization of academic voice, and reduced self-regulation when AI use is insufficiently guided. This paper adds to existing debates on AI, academic literacy, and integrity in higher education by conceptualizing multimodal AI as a mediational resource rather than a surrogate author

Fields of Study & Subjects

Writing and Handwriting Education Discourse Analysis in Language Studies Artificial Intelligence in Healthcare and Education

Publication details

Journal / source
Arab World English Journal
Publisher
AWEJ Publishing
Issue
3
Pages
390–403
Publication date
23 January 2026