The augmented scholar: how Generative AI usage influences research self-efficacy among Chinese doctoral students

This study examines the relationship between Generative AI (GenAI) usage and research self-efficacy (RSE) among Chinese doctoral students, a topic of growing importance as AI technologies become embedded in academic practices. Using a quantitative, cross-sectional survey design, data were collected via a survey from 587 doctoral students across STEM and humanities and social sciences (HSS) disciplines. Hierarchical multiple regression and moderation analyses (PROCESS macro) were employed to examine predictive relationships and interaction effects. Results revealed that both the frequency and purpose of GenAI use significantly predicted RSE, with an augmentative approach – using AI as a critical thinking scaffold – emerging as the strongest positive predictor. Furthermore, perceived supervisor support and academic discipline moderated this relationship, with supportive supervision amplifying the benefits of augmentative use, and STEM students demonstrating a stronger link between frequency of use and RSE than their HSS counterparts. The findings underscore the importance of fostering critical AI literacy and providing supervisory guidance to maximize the psychological and scholarly benefits of GenAI integration. This study contributes to doctoral education research by empirically demonstrating how augmentative GenAI usage serves as a psychological scaffold that enhances research self-efficacy, and by identifying supervisor support and disciplinary context as key moderating factors in this relationship.