Artificial intelligence and intimacy development in university students: A gender-based systematic review and meta-analysis

Abstract

Background: Artificial intelligence (AI) increasingly shapes university students’ interpersonal communication and emotional engagement, influencing intimacy development in digitally mediated relationships. However, gender-based differences and contextual moderators remain insufficiently synthesized. Objective: This study systematically reviewed and meta-analyzed empirical evidence on the association between AI-mediated communication and intimacy development among university students, with particular attention to gender differences and moderating factors. Methods: Thirty-two studies involving 18,457 participants were included. A comprehensive search of major academic databases was conducted from January 2021 to December 2025. Study selection, data extraction, and quality assessment were independently performed by two reviewers. Random-effects meta-analyses, subgroup analyses, meta-regression, sensitivity tests, and publication bias assessments were conducted. Results: The pooled analysis demonstrated a moderate positive association between AI-mediated communication and intimacy development (r = .27, 95% CI [.25, .29]). Stronger effects were observed among female students (r = .33) than male students (r = .24). Age, cultural context, AI system type, duration of use, and platform characteristics significantly moderated effect sizes. Findings were robust across analytical models, with minimal evidence of publication bias. Conclusions: AI-mediated communication plays a meaningful role in shaping intimacy among university students, with notable gender and contextual variations. Balanced, emotionally responsive AI use appears to support relational development, whereas excessive reliance may pose psychosocial risks. These findings highlight the importance of gender-sensitive digital literacy, ethical AI design, and targeted interventions in higher education.