Paper Type

Complete

Abstract

This study systematically reviews research on trust and distrust in educational generative artificial intelligence (GenAI). Using PRISMA 2020 guidelines, we synthesize 53 empirical studies from Scopus and Web of Science to examine how trust and distrust are theorized, operationalized, and empirically examined in GenAI-enabled educational contexts. Findings show that existing research largely treats trust as an adoption mechanism within TAM/UTAUT-style models, while distrust is often underdeveloped or reduced to low trust. The synthesis identifies four ecosystem dimensions shaping trust formation: AI characteristics, user characteristics, task characteristics, and environmental context. We further develop a trust calibration perspective to explain when learners’ confidence aligns—or misaligns—with GenAI’s actual reliability. This review contributes a conceptual framework that positions trust and distrust as coexisting, context-dependent mechanisms and outlines future research directions beyond adoption-focused models toward more responsible educational GenAI use.

Paper Number

1361

Comments

SIG ED

Download

Share

COinS
 
Aug 15th, 12:00 AM

Trust and Distrust in Educational Generative Artificial Intelligence: A Systematic Review and Thematic Synthesis

This study systematically reviews research on trust and distrust in educational generative artificial intelligence (GenAI). Using PRISMA 2020 guidelines, we synthesize 53 empirical studies from Scopus and Web of Science to examine how trust and distrust are theorized, operationalized, and empirically examined in GenAI-enabled educational contexts. Findings show that existing research largely treats trust as an adoption mechanism within TAM/UTAUT-style models, while distrust is often underdeveloped or reduced to low trust. The synthesis identifies four ecosystem dimensions shaping trust formation: AI characteristics, user characteristics, task characteristics, and environmental context. We further develop a trust calibration perspective to explain when learners’ confidence aligns—or misaligns—with GenAI’s actual reliability. This review contributes a conceptual framework that positions trust and distrust as coexisting, context-dependent mechanisms and outlines future research directions beyond adoption-focused models toward more responsible educational GenAI use.