Rethinking Citation of AI Sources in Student-AI Collaboration within HCI Design Education (2506.08467v2)

Abstract: The growing integration of AI tools in student design projects presents an unresolved challenge in HCI education: how should AI-generated content be cited and documented? Traditional citation frameworks -- grounded in credibility, retrievability, and authorship -- struggle to accommodate the dynamic and ephemeral nature of AI outputs. In this paper, we examine how undergraduate students in a UX design course approached AI usage and citation when given the freedom to integrate generative tools into their design process. Through qualitative analysis of 35 team projects and reflections from 175 students, we identify varied citation practices ranging from formal attribution to indirect or absent acknowledgment. These inconsistencies reveal gaps in existing frameworks and raise questions about authorship, assessment, and pedagogical transparency. We argue for rethinking AI citation as a reflective and pedagogical practice; one that supports metacognitive engagement by prompting students to critically evaluate how and why they used AI throughout the design process. We propose alternative strategies -- such as AI contribution statements and process-aware citation models that better align with the iterative and reflective nature of design education. This work invites educators to reconsider how citation practices can support meaningful student--AI collaboration.