AI Dependency and Student Well-Being: The Moderating Role of Personality Traits in Social, Psychological, and Educational Outcomes

Abstract

Artificial intelligence (AI) tools like ChatGPT, Copilot, and Gemini have significantly transformed how university students learn, write papers, conduct research, and manage academic stress. While these benefits are evident, concerns about over-reliance on AI and its social, psychological (self-esteem), and educational impacts remain. This study examined the relationship between AI dependency and the social, psychological, and educational challenges faced by university students. It also explored the Big Five personality traits as moderators to assess how they influence the relationship between AI dependency and students’ challenges in Pakistan. The quantitative phase (N = 350) used survey data, including the GenAI Dependency Scale, the Big Five Inventory, the Social Connectedness Scale, the Rosenberg Self-Esteem Scale, and the Educational Stress Scale for Adolescents. Findings indicated that AI dependency was negatively associated with social connectedness (r = -.219, p < .001) and positively associated with educational stress (r = .325, p < .001), with a less significant relation to self-esteem. Additionally, agreeableness moderated the relationship between AI and social connectedness, while multiple personality traits moderated the relationship between AI and self-esteem. The results highlight AI's dual role as an academic aid and a potential source of disengagement. Recommendations are provided for students, teachers, counselors, universities, and policymakers to promote responsible, balanced, and psychologically healthy use of AI in higher education.