Students’ strategies for revising AI-assisted writing: towards a conceptual framework

As generative artificial intelligence (GenAI) becomes integrated into academic writing, understanding how second language (L2) English students revise AI-assisted writing remains underexplored; therefore, this inductive content analysis explores students’ revision strategies for academic writing to propose a conceptual framework . The study employs an open-ended free listing task with 100 students, who listed their five major strategies for revising AI-assisted writing assignments. Additionally, semi-structured interviews were conducted with 10 students experienced in revising AI-assisted writing. The interview transcripts were open-coded to develop categories. Meanwhile, the free-listing responses were directly categorized because participants listed strategies as explicit codes. The researcher then compared, consolidated, and synthesized the strategy categories into broader themes to form a conceptual framework. The analysis found key revision strategies like incorporating personal ideas and expressions, simplifying the language, consulting additional sources, paraphrasing and rewriting content, integrating personal experience, contextualizing AI content, and cross-checking AI information. It suggested that L2 English students primarily personalize and improve AI-assisted writing rather than passively accepting AI-generated text, while also adapting, evaluating, and responsibly managing AI-generated content. This final synthesis yielded six overarching themes: supportive utilization, ethical engagement, creative reform, critical evaluation, contextual revision, and functional adjustment. The proposed framework with six themes serves as an empirically grounded taxonomy of student revision strategies in AI-assisted writing.