Large Language Models and Emotional Regulation of University Students in Peshawar, Khyber Pakhtunkhwa

Abstract

In the digital age, LLMs have significantly transformed students’ academic practices and impact emotional experiences. The present study aimed to investigate the impact of instrumental and relational dependency on emotion regulation strategies among university students. A quantitative, cross-sectional research design was employed. The sample consisted of 380 university students, recruited through convenience sampling from multiple universities. Participants completed a demographic information sheet, the Dual-Dimensional Scale of Instrumental and Relationship Dependency on Large Language Models (LLM-D12), and the Emotion Regulation Questionnaire (ERQ), which measures habitual use of cognitive reappraisal and expressive suppression. Data were analyzed using descriptive statistics, multivariate regression, and independent-samples t-tests using SPSS version 23. Findings revealed that instrumental dependency does not significantly predict cognitive reappraisal or expressive suppression. Additionally, relational dependency showed a significant negative prediction effect on cognitive reappraisal and expressive suppression, suggesting that greater relationship dependency was associated with lower use of these emotion regulation strategies. Furthermore, non-significant gender differences were found across instrumental dependency, relational dependency, cognitive reappraisal, and expressive suppression. The findings underscore the need for balanced, responsible, and psychologically informed approaches to AI integration that support academic performance while preserving students’ emotional growth and independence in an increasingly AI-driven educational environment.