Relationship between artificial intelligence tool usage experience and academic stress among college students: Mediating role of loneliness and moderating role of academic self-efficacy

Overview

Survey study examining associations between college students' artificial intelligence (AI) tool usage experience and academic stress, with tests of loneliness as a mediator and academic self-efficacy as a moderator. Sample comprised 624 university students. Self-report instruments used were the AI Tool Usage Experience Scale, UCLA Loneliness Scale, Academic Stress Scale, and Academic Self-Efficacy Scale. The study reports direct and indirect pathways linking AI tool usage experience to academic stress and a moderation effect of academic self-efficacy on the first stage of the mediation pathway.

Methods and approach

A survey study of 624 university students using self-report instruments (the scales listed in the abstract). Mediation and moderation analyses were conducted to evaluate whether loneliness mediated the relationship between AI tool usage experience and academic stress, and whether academic self-efficacy moderated the first stage of that mediation pathway. No additional methodological details (e.g., statistical controls, specific modeling techniques) are reported in the abstract.

Results

AI tool usage experience was positively associated with academic stress. Loneliness partially mediated this relationship, indicating that part of the association between AI tool usage experience and academic stress operated via increased loneliness. Academic self-efficacy moderated the first stage of the mediation: the positive association between AI tool usage experience and loneliness was reported as stronger for students with higher academic self-efficacy and weaker for students with lower academic self-efficacy.

Implications

Findings indicate that AI tool usage experience relates to higher academic stress both directly and indirectly through increased loneliness, highlighting a psychosocial pathway relevant to technology integration in higher education. The moderation by academic self-efficacy suggests heterogeneity in psychological responses to AI tool use, with stronger links to loneliness among students reporting higher academic self-efficacy. Future research should employ designs that permit causal inference, report analytic specifications and covariate handling explicitly, and examine boundary conditions and mechanisms underlying the observed moderation and mediation effects.