Empowerment or disempowerment: the impact of GenAI use on graduate students' research innovation capability

Purpose This study aims to examine how generative artificial intelligence (GenAI) use influences graduate students' research innovation capacity in the Chinese context, with particular attention to the mediating role of research self-efficacy, the moderating effects of peer support and institutional support and heterogeneity across levels of academic training and disciplinary fields. Design/methodology/approach Survey data were collected from 747 graduate students at a comprehensive research-oriented university in China. Structural equation modeling was employed to examine the relationships among GenAI use, graduate students' research self-efficacy and research innovation capacity. Moderated mediation analyses were conducted to assess the moderating effects, while hierarchical regression analyses were used to explore heterogeneity. Findings The results suggest that GenAI use is positively associated with graduate students' research innovation capacity, research self-efficacy partially mediates this relationship. Moreover, peer support and institutional support significantly strengthen both the direct and indirect effects of GenAI use on research innovation capacity. The empowering effects of GenAI use are more pronounced among doctors and students in the humanities and social sciences. Originality/value This study integrates stimulus–organism–response theory with the perspective of professional capital to develop a framework that conceptualizes GenAI use as a form of digital stimulus embedded in professional capital. The findings elucidates how GenAI use influences the research innovation capacity of graduate student communities as a manifestation of human capital, highlights the heterogeneity of these effects and explains how peer support and institutional support, as forms of social capital, moderate these relationships.