<![CDATA[Academic writing in English can be particularly challenging for EFL graduate students because it requires grammatical accuracy, coherent organization of ideas, and effective argumentation. Generative Artificial Intelligence (GenAI) has emerged as a potential tool to support this process, yet research on its acceptance has largely relied on quantitative surveys that provide limited insight into students’ experiences. This study explored three EFL graduate students’ perceptions of GenAI in supporting English academic writing through the Technology Acceptance Model (TAM). Specifically, it explored perceived usefulness and ease of use, attitudes toward GenAI, and behavioral intention to use it for academic writing. A qualitative study was conducted with three Master’s students who regularly used GenAI for academic writing. Data were collected through semi-structured interviews and analyzed thematically. The findings suggest that, among these participants, perceived usefulness varied across writing processes, with GenAI serving corrective, cohesive, and generative functions. Perceived ease of use was influenced not only by the accessibility of the tool but also by participants’ ability to formulate effective prompts. Positive attitudes were associated with greater autonomy, whereas negative attitudes emerged when AI-generated suggestions conflicted with participants’ academic judgment. Despite these concerns, the three participants intended to continue using GenAI while maintaining boundaries through teacher feedback and crosschecking AI-generated suggestions. Within this context, the findings suggest that GenAI use may be experienced as bounded and negotiated rather than as straightforward acceptance. It is recommended that EFL writing instruction integrate GenAI alongside human feedback and encourage students to critically evaluate AI-generated output. ]]>
