The bright and dark sides of AI in learning: How AI ‐driven challenge and hindrance technostressors affect undergraduate students' academic performance

Abstract The rapid integration of artificial intelligence (AI) into university learning environments has fundamentally reshaped how undergraduate students access information, complete academic tasks and engage with course content. While AI technologies promise enhanced learning efficiency and personalised support, they also introduce new forms of technostress that may either facilitate or hinder academic performance. Drawing on the challenge–hindrance stressor framework and the Job Demands–Resources (JD‐R) model, this study distinguishes between AI‐driven challenge technostressors (AICS) and AI‐driven hindrance technostressors (AIHS) to examine their differential effects on undergraduate students' academic performance. We further investigate how traditional support structures and peer support structures moderate these relationships. Using longitudinal survey data collected from 391 undergraduate students with AI‐supported learning experience, the results reveal that AICS positively influences academic performance, whereas AIHS exerts a negative effect. Moreover, traditional support structures buffer the negative impact of AIHS but weaken the positive effect of AICS. In contrast, peer support structures strengthen the positive relationship between AICS and academic performance but do not significantly mitigate the negative effect of AIHS. This study contributes to the technostress literature by introducing and empirically validating the dualistic nature of AI‐driven technostressors in educational settings. By extending the challenge–hindrance framework to AI‐enhanced learning and integrating contextual support mechanisms within the JD‐R model, the findings offer nuanced theoretical insights and practical guidance for institutions seeking to optimise AI implementation while promoting sustainable student performance.