When leadership becomes epistemic: paradoxes of educational leadership in AI-mediated systems

Abstract

The growing presence of artificial intelligence in education does not merely introduce new tools; it fundamentally shifts the context that educational leadership must confront: the logic of information scarcity and coordination is replaced by interpretive overload and uncertainty regarding knowledge legitimation. The study argues that this shift makes the leadership of AI-mediated educational systems fundamentally epistemic and, as a result, structurally prone to paradox. Established (instructional, transformational, distributed) and technocratic models are insufficient to capture this condition on their own, as each assumes the stability of knowledge legitimation. By integrating paradox theory and epistemic leadership, the study identifies six structural leadership paradoxes, which are organized into three clusters (pedagogical, epistemic, and temporal-organizational). The common structural consequences of these paradoxes—interpretive overload, the compression of time for reflection, and increasing psychological strain—call for two focal, mutually dependent leadership capacities: reflective sensemaking as an epistemic operation, and digital resilience as its enabling condition. The study concludes that the focus of educational leadership in the AI era is shifting from technological control toward epistemic orientation, and it advances a set of testable propositions through which this claim—and the proposed framework as a whole—can be empirically examined.