The Psychological Effects of AI Companions on Loneliness and Perceived Social Connection Among Young Adults

Abstract

Artificial intelligence (AI) companions are becoming more popular as conversational and emotional supports, but their association with loneliness and social connection among young adults in South Asia has not been extensively studied. The present study investigated the relationship between using AI companions, loneliness, and perceived social connection in young adults. The design employed was quantitative cross-sectional, involving 280 participants aged 18-30 years who have experienced using AI conversational companions like ChatGPT, Replika, or Character. The participants were AI who were recruited by purposive sampling. Loneliness was assessed using the UCLA Loneliness Scale, perceived social connection with the Social Connectedness Scale–Revised (SCS-R) and AI companion engagement with interaction frequency, emotional reliance, and perceived companionship items. Descriptive statistics, Pearson correlation analysis, and hierarchical multiple regression (controlling for age, gender, and educational level) were used for data analysis. The results showed that greater loneliness was correlated with greater AI companion engagement (β = .38), while greater perceived social connection was correlated with less engagement (β = −.33); the latter correlated with greater engagement at the component level, whereas the former did not. The design was cross-sectional, giving the results that are compatible with the postulates of both displacement and compensation. They recommend that the quality of and emotional function of the use of AI should also be taken into account in student support and digital wellbeing campaigns, and not just the frequency of use.