''It's Safer to Say I Don't Use It'': Exploring ESL Students' Self-Perceptions as AI-Assisted Writers

Many English-as-a-Second-Language (ESL) students have begun to use generative artificial intelligence (AI) tools to improve their writing artifacts. This introduces both new opportunities and challenges in how they develop, express, and perceive themselves as writers. Drawing on 30 semi-structured interviews with ESL students with experience using generative AI tools in their English writing, this study explores how they incorporate these tools into their writing process, and how doing so reshapes their self-perceptions as writers. Our findings reveal that ESL students strategically use generative AI tools not only to improve the correctness of their language but also to navigate rhetorical and cultural nuances in English writing across different levels. With the language barrier alleviated, they reflected critically on what the use of generative AI means for their writing motivation, personal expression, and skill development as writers. Yet, they found themselves often caught in between conflicting rules and social norms regarding generative AI usage, which creates a ''non-win'' situation where ESL students face potential judgment regardless of their generative AI usage choices. Drawing on these findings, we conclude by offering recommendations for educators to help ESL students engage with writing critically and confidently, and to support their continuous development as AI-assisted writers.