Revisiting Doctoral Liminality in the Era of Artificial Intelligence (AI) and Digital Scholarship: Implications for Research Integrity and Supervision

Abstract

Doctoral education is undergoing profound transformation due to the rapid integration of artificial intelligence (AI) and digital technologies; however, this shift has introduced significant challenges related to research competence, supervision, and research integrity. Despite the growing use of AI-assisted research tools, there remains limited empirical understanding of how these technologies interact with doctoral students’ transitional (liminal) experiences and institutional support systems to influence ethical research outcomes. In particular, concerns have emerged regarding the potential for AI misuse, weakened critical thinking, and gaps in supervisory guidance, raising critical questions about how doctoral training systems can effectively sustain research quality and integrity in the digital era. This study addresses this gap by providing a rigorous empirical and theoretical examination of doctoral liminality within the context of AI and digital scholarship. Specifically, it investigates how doctoral liminality, digital research environments, AI tool usage, and supervisory support influence research competence and responsible research conduct. Adopting a quantitative cross-sectional design, data were collected from 320 doctoral students across selected universities in Kenya and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). The results reveal a robust and well-fitting model with strong explanatory power (R² = 0.62 for research competence; R² = 0.55 for research integrity). AI tool usage emerged as the strongest predictor of research competence (β = 0.36, p < 0.001), followed by doctoral liminality, digital research environment, and supervisory support. Research competence, in turn, exhibited a strong positive effect on responsible research conduct (β = 0.74, p < 0.001), confirming its central role as a pathway to ethical research behavior. Furthermore, research competence significantly mediates the relationship between technological and developmental factors and research integrity, while supervisory support plays both a direct and moderating role in enhancing the effective and ethical use of AI tools. By integrating technological, institutional, and behavioral perspectives, this study advances the conceptualization of doctoral liminality in the digital era. It contributes to emerging scholarship on AI-enabled research and provides policy-relevant insights for strengthening doctoral supervision, embedding AI governance, and promoting competence-based approaches to research integrity in increasingly digitized academic environments.