Educational institutions are undergoing transition, as students, and educators increasingly adopt generative AI. Therefore, it is important to examine stakeholders' perspective and their interpretations of safe, ethical, and responsible AI. Undertaking a micro-ethnographic case study, this study contributes to the literature in three ways. Along with an empirical study, it highlights students’ perspectives that remains underrepresented. The study also presents empirical evidence of a ‘distributed-constitutive pathway’ of imaginary formation, based on the human-GenAI relationship. A discourse thematic analysis identified three sociotechnical imaginaries – about technology, human-GenAI relationship and education. A narrative about the environmental cost of GenAI was also identified. The first and second imaginary shape institutional policy and practice and operate on the global technological movement. The third imaginary is aligned with policy but resisted by the other two imaginaries. Key findings indicate that the technological movement manifests as a sense of technological inevitability. It causes policy to be reactive and creates gaps between policy and the lived experience of students. A promising bridge between is found to be AI literacy. Together, these suggest that the safe, responsible, and ethical use of GenAI is a contested, negotiated, complex, and coproduced process within the institution.