In this paper, I examine how generative AI reshapes learner authorship in higher education through the lens of Donna Harraway’s Cyborg Theory and Bruno Latour’s Actor–Network Theory. Both theories support the notion of ‘hybridity’ in the human body and society, through which the boundaries between human and non-human agencies are questioned. Applying the theories, this study proposes to rethink of the relationship between learners and their learning tools, particularly Generative AI tools. AI technologies develop rapidly, and so does the education sector’s response to them. In particular, scholarship has been shifting from policing AI misuse to reconsidering assessment validity in AI-rich environments, emphasising the distinction between critical and uncritical engagement rather than AI use versus non-use. The current study is motivated from the need of recognising learner’s critical engagement with Ais through redesign of assessment of the Korean Stage 3 module; a dual assessment structure combining written tasks with oral presentations, alongside students’ reflection on their use of generative AI tools. While the oral presentation component worked as the opportunity for students to claim their authorship on their work and for the teacher as “AI-vulnerable” assessment tool (Roe et al, 2026), their reflection on AI use provides them with an opportunity to evaluate their own relationship with the AI tools. While students are eager to frame AI tools as supportive devices and thereby reinstate the hierarchical relationship, a close reading of their reflections reveals their anxious urge to assert that the final product is their own. Such anxiety stems from their denial of the closely intertwined relationship between themselves and the AI tools. Through reconceptualising assessment as a collaborative human–AI practice and calling for institutional cultures that support flexibility in their use of AI tools, the study proposes redefining the student authorship in the age of generative AI.