Text Production Versus Text Orchestration: The Efficacy of ChatGPT in EFL Academic Writing

Abstract

EFL writing in educational settings has been greatly impacted by AI tools as it has transitioned from  independent text production to text orchestration, a process in which students utilize AI to organize their writing, improve language use, and generate ideas. Integration of new technologies in education calls for careful evaluation of its pros and cons. This quasi-experimental study analyzes how EFL learners’ text production and writing quality are affected by ChatGPT-mediated writing. The study sample of 102 intermediate EFL students from Qassim University were assigned in equal numbers to two groups, viz., control (CG) and experimental (EG). Over a period of ten weeks, participants undertook academic essay writing based on a standardized assessment system with CG using conventional educational resources and EG using ChatGPT for idea generation, linguistic inputs, and structural organization. Writing performance was evaluated using the IELTS analytic rubric that assesses four parameters of writing output. Writing performance of both groups and role of ChatGPT in aiding the EG were examined using one-way ANOVA, independent samples t-tests, and descriptive statistics using IBM SPSS (Version 26). Results indicated higher mean scores in the EG than CG across all categories with statistically significant differences across groups (t = 4.86–6.18, p < 0.001; F = 18.72, p < 0.001). The study concluded that AI-mediated writing significantly improves writing organization, linguistic accuracy, and vocabulary use, making a case for transitioning EFL learners’ writing from independent text production to AI-assisted text orchestration. The study has implications for many stakeholders including teachers, edtech developers, and learners.