<![CDATA[The integration of generative artificial intelligence into higher education has transformed the ways postgraduate students begin and advance their academic work. However, few studies have explored how AI influences epistemic authority in the development of thesis proposals. This study examines how postgraduate students establish and prioritize AI, peers, and supervisors as epistemic authorities within an AI-mediated academic environment. Using a qualitative interpretative methodology, we gathered data from 105 graduate students through questionnaires and semi-structured interviews. Reflexive thematic analysis unveiled a stratified epistemic hierarchy. AI was seen as an important authority because it could help people think more clearly and generate new ideas, but it had to be checked. Peers acted as relational authorities, giving emotional support and dialogic assistance. Supervisors maintained the highest epistemic status as institutional authorities owing to their evaluative authority, responsibility, and symbolic capital within academic frameworks. AI reduced uncertainty and increased perceived control in the early stages of proposal creation, without undermining supervisory legitimacy. Authority remained grounded in institutional acknowledgment rather than in the rapid dissemination of knowledge. The findings enhance AI-in-education studies by shifting the emphasis from usage frequency to authority configuration, offering insights for AI governance and postgraduate supervision.]]>
Copyrights © 2026
