Abstract

Abstract

This study aims to examine, through a multidimensional approach, how artificial intelligence technologies are transforming research paradigms, data production processes, and knowledge construction methods in educational sciences. The research seeks to fill a significant gap in the literature by considering AI not only as a technical tool but also as a transformative element reshaping the methodological and epistemological foundations of educational research. In this respect, the study goes beyond the tool-oriented approaches seen in most existing research, comprehensively analyzing the impact of AI on scientific knowledge production processes at the theoretical and methodological levels. Furthermore, the study offers a critical perspective questioning the extent to which educational sciences are prepared for this transformation, given the increasing prevalence of data-driven research approaches. The study is designed using an integrative review approach based on systematic document analysis. In this context, academic studies obtained from international databases were selected according to specific inclusion and exclusion criteria and grouped under a single dataset. Bibliometric analysis techniques were used to reveal general trends in the literature, publication densities, citation networks, and thematic clustering. In addition, the conceptual foundations, methodological implications, and epistemological consequences of AI integration were examined in depth through thematic content analysis performed on the same dataset. The findings obtained during the analysis process were systematically coded, inter-thematic relationships were revealed, and interpreted. This two-way analysis approach both makes visible the macro-level development dynamics of the field and provides qualitative depth at the micro-level. The findings show that AI-supported methods significantly increase data processing capacity in educational research, enable the analysis of large and multidimensional datasets, and bring speed, accuracy, and diversity to research processes. In particular, methods such as machine learning, natural language processing, learning analytics, and educational data mining were found to produce more comprehensive and predictive results, going beyond traditional analysis techniques.