ARTIFICIAL INTELLIGENCE-BASED WRITING ASSISTANCE FOR ENGLISH LANGUAGE LEARNERS WITH SPEECH IMPAIRMENTS

Abstract

Inclusive education emphasizes the right of every learner to access meaningful, high-quality instruction, including students with speech impairments, whose writing development is often hindered by limited opportunities for oral expression, feedback, and interaction. In many special-school settings, teachers struggle to provide individualized writing support due to time constraints and diverse learner needs. Responding to this gap, the present study investigates the extent to which three AI-assisted writing applications, Grammarly, ChatGPT, and QuillBot, can support the development of English writing skills among students with speech impairments in Indonesia. Using a mixed-methods approach, the research involved 60 learners who participated in a series of guided writing tasks. Quantitative data were obtained from pre- and post-test assessments, while qualitative insights were gathered through semi-structured interviews with students and teachers as well as classroom observations. The findings demonstrate consistent improvement across key writing dimensions. Students showed better command of grammar and vocabulary, and their written texts displayed clearer organization and more coherent argumentation. Teachers also reported increased confidence and engagement among learners, noting that AI tools provided immediate feedback that would otherwise be difficult to deliver during regular instruction. While the tools did not replace pedagogical guidance, they functioned as effective scaffolds that helped students revise their work more independently. Overall, the study provides empirical support for the use of AI-based writing tools in special education settings. It highlights their potential to enhance accessibility, reduce instructional barriers, and strengthen writing outcomes for learners with speech impairments.