The Role of Interaction with Artificial Intelligence Technologies in Shaping the Psychological Wellbeing of University Students: A Systematic Review and Meta-Analysis
Main Article Content
Abstract
University students face disproportionately high rates of mental health disorders yet remain chronically underserved by conventional psychological services. This systematic review synthesised empirical evidence on the role of AI technology interaction in shaping university students' psychological wellbeing. Seven electronic databases were searched with no date restrictions, following PICOS eligibility criteria and PRISMA 2020 reporting guidelines. Methodological quality was appraised using the Mixed Methods Appraisal Tool 2018 (MMAT), and the overall certainty of evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) framework. Findings were synthesised through narrative thematic synthesis, supplemented by an exploratory random-effects quantitative synthesis where sufficient homogeneous data were available. Forty studies (N = 9,692) were included. Cognitive-behavioural therapy-based AI chatbot interventions produced significant reductions in depression and loneliness (Hedges' g = −0.43; 95% CI: −0.62, −0.24), while adaptive learning platforms improved self-efficacy and life satisfaction. Conversely, excessive or unstructured AI use was associated with technostress (g = 0.44), AI dependency (g = 0.38), social isolation, and academic burnout. AI literacy, satisfaction of basic psychological needs, and purposive tool design emerged as the most consistent protective moderators. AI interaction exerts clinically meaningful but contingent effects on student psychological wellbeing: benefits are realised only when tools are therapeutically grounded, institutionally scaffolded, and positioned as complements to — rather than substitutes for — human care. Longitudinal, cross-cultural research employing standardised outcome measures is urgently needed to advance causal understanding of AI–wellbeing dynamics in higher education.