An ecological systems analysis of university students' ChatGPT use in higher education

Abstract

Educational technology research has largely conceptualised students' ChatGPT adoption through individual-level cognitive or integrity frameworks, leaving multilevel systemic conditions undertheorised. This study applies Bronfenbrenner's ecological systems theory (EST) to examine how students at a research-intensive university in Istanbul, Turkey, comprising native Turkish speakers, international students in language preparatory programmes and proficient multilinguals, negotiate ChatGPT use across nested ecological contexts. Drawing on 43 semistructured interviews analysed through reflexive thematic analysis, five themes spanning EST's ecological levels were identified. Key findings reveal ChatGPT functioning as a relational and compensatory resource, cross-level cascades from policy ambiguity to microsystem anxiety, macrosystem linguistic ideologies producing inequitable consequences for multilingual students, and chronosystem identity transformations under policy instability. The study contributes an ecological framework with contextually grounded principles that may inform understanding of AI adoption in comparable multilingual, internationalising university contexts and points to multilevel interventions addressing policy coherence, resource equity and pedagogical transformation.