This study investigates qualitative changes in the identity transformation of advanced Chinese learners of Korean, particularly for academic purposes, through the internalization of generative artificial intelligence (GAI)-based feedback. Utilizing a narrative inquiry methodology, three rounds of semi-structured online and offline interviews were conducted with two Chinese graduate students enrolled at a Korean university. Their experiences with ChatGPT-based feedback were analyzed within a temporal and contextual framework. The findings reveal that the learners internalized GAI feedback not merely as a tool for error correction but also as a means to strategically monitor and regulate their linguistic output through self-regulated practices. Their self-regulatory competence expanded to encompass discourse structuring, thought flow organization, and formation of audience-oriented expression strategies, illustrating their transition into legitimate academic performers within the academic discourse community. This study empirically demonstrates that GAIbased feedback contributes not only to linguistic proficiency but also to the development of academic self-expression and ontological identity formation. Furthermore, it discusses the potential of GAI feedback as a meaningful pedagogical mediator that supports the process of linguistic selfing and the construction of subjectivity among foreign language learners for academic purposes.