Linguistic mediation of artificial intelligence in academic writing and reading in higher education: advantages, challenges, and pedagogical perspectives

Abstract

This study examines the advantages and challenges posed by the linguistic mediation of Artificial Intelligence (AI) in academic writing and reading within higher education. Based on a systematic documentary review of 20 academic sources published between 2020 and 2025 in databases such as Scopus, Web of Science, and Google Scholar, thematic patterns were identified regarding the linguistic functions, benefits, and limitations of different AI tools. Findings show that applications such as Grammarly, ProWritingAid, Hemingway Editor, Writefull, Read & Write, Vocaro, and VoiceThread enhance grammatical correction, textual coherence, lexical fluency, reading comprehension, and oral feedback, contributing to the development of communicative competence. However, significant risks are also noted, including the loss of authorship and authentic voice, the homogenization of linguistic style, cognitive dependency, and bias reproduction. The study concludes that the responsible use of AI requires digital literacy, ethical training, and pedagogical strategies that foster critical and reflective learning, ensuring that technology complements rather than replaces human capabilities.