Speaker
Description
The rapid adoption of generative artificial intelligence (AI) has challenged traditional notions of authorship and academic integrity in university writing. While previous research has primarily examined students' attitudes toward AI or the ethical implications of its use, less attention has been paid to how students negotiate authorship boundaries during AI-assisted writing. This study investigated the decision-making processes through which university students determined ownership, responsibility, and disclosure of AI-assisted writing. The study involved 76 undergraduate students enrolled in two academic writing classes. All participants completed an authentic writing task using AI as they normally would, and 18 purposively selected students participated in stimulated-recall interviews based on their writing processes. Data from writing drafts, revision histories, AI interaction logs, interview transcripts, and institutional AI policy documents were analyzed using constructivist grounded theory methods. The findings revealed that students viewed authorship not as a binary distinction between human- and AI-generated writing but as a negotiated judgment shaped by the purpose of AI use, the extent of human revision, the perceived originality of the final text, and their sense of intellectual responsibility. AI assistance was generally considered acceptable for surface-level support, such as grammar, vocabulary, and outlining, but less acceptable when it substantially contributed to idea generation or drafting. The analysis resulted in a conceptual framework of AI authorship-boundary negotiation, illustrating how students claimed authorship through interconnected judgments of ownership, human contribution, revision depth, ethical responsibility, and awareness of institutional expectations. The framework advances current understandings of AI-assisted writing and offers practical guidance for university AI policies, assessment design, and academic integrity education.
