<p>The increasing integration of ChatGPT into higher education offers students a plethora of benefits, but also raises controversial debates regarding academic integrity, particularly in the context of plagiarism using ChatGPT. This issue has been discussed in both international and domestic research domains, yet the latter still severely lacks relevant studies, and the former has produced mixed, contextual findings. Therefore, this study aims to fill the gaps by investigating university students’ perceptions of plagiarism in the use of ChatGPT through the dual-cluster model of traditional plagiarism and AI-giarism. The data were drawn from 183 questionnaire participants and 15 interviewees at a private university in the Mekong Delta, Vietnam. The descriptive findings indicate that most participants had a consistent perception of the traditional forms of plagiarism. Conversely, descriptive findings from the AI-giarism cluster signal ambivalence in students’ perception. The qualitative findings were further used to explain major patterns and themes found in the questionnaire data. Qualitative findings demonstrate that students were strongly opinionated about more obvious forms of plagiarism, knowledgeable about academic integrity issues besides plagiarism, and conflicted about the ambiguity of some AI-giarism scenarios. Based on the findings, this study recommends that institutions should construct more consistent, balanced, and fairer policies to tackle the issue of plagiarism in the use of ChatGPT among students. Furthermore, some recommendations are given to educators on how to address plagiarism issues more effectively in their classrooms.</p>