Using Diffit AI for Text Adaptation to Enhance Writing Skills The Case of Third-Year English Students at The University of Ghardaïa

Abstract

This study investigates the effectiveness of using Diffit AI for text adaptation to enhance the writing skills of Third-Year Licence students at the University of Ghardaïa. Writing remains a persistent challenge for English as a Foreign Language (EFL) learners, who frequently struggle with vocabulary acquisition, syntactic structures, and rhetorical organization. To address these individual learner needs, this research explores the integration of Artificial Intelligence (AI) as a personalized scaffolding tool. A mixed-methods research design was adopted to collect and analyze both quantitative and qualitative data. The quantitative component utilized pre-tests and post-tests to measure immediate improvements in the participants' writing performance, specifically focusing on grammar, vocabulary, and paragraph organization. The qualitative component employed questionnaires and short interviews to explore students' attitudes, experiences, and perceptions regarding the interactive use of Diffit AI. The research operates under the hypotheses that AI-driven text adaptation positively influences students' linguistic accuracy and that learners maintain positive attitudes toward the tool due to increased engagement and reduced text frustration. While the study is bounded by specific limitations, including a relatively small sample size, a short intervention duration, and potential technical or internet constraints, the findings provide valuable pedagogical insights for educators at the University of Ghardaïa and similar institutions on integrating generative AI tools to support differentiated instruction and writing development.