Abstract

This research explores how algorithm‑driven content exposure on mainstream social media platforms predicts subjective loneliness among Generation Z young adults in the United States, and examines whether weak‑tie social connection acts as a mediating variable in this relationship. Scholars have long noted the connectivity paradox of social media: although digital platforms greatly expand people’s opportunities for social contact, heavy social media engagement does not necessarily reduce subjective loneliness, and may even intensify feelings of social isolation. Existing research has mostly focused on total screen time or active versus passive browsing behaviors, while fewer studies specifically separate algorithmic content exposure from general social media use. Algorithmic recommendation systems prioritize user retention by delivering personalized content, which shapes the types of social relationships users develop online. Many prior studies treat social media use as a monolithic activity, failing to acknowledge that content curated automatically by algorithms differs fundamentally from content selected through intentional subscription and friend‑following. This oversight creates gaps in understanding how platform architecture, rather than merely usage duration, shapes social psychological outcomes. This study hypothesizes that greater exposure to algorithmically curated social media content positively predicts weak‑tie social connection, and that expanded weak‑tie connections further mediate the association between algorithmic exposure and subjective loneliness. Data were collected through an online cross‑sectional questionnaire distributed to United States participants aged between 18 and 26 years old. The final valid sample contained 714 respondents. Measurement scales adopted internationally validated instruments including the UCLA Loneliness Scale and the Weak‑Tie Social Connection Scale. Statistical analysis was conducted using bootstrapping mediation procedures. The analytical results show that algorithmic content exposure has a significant positive association with weak‑tie social connection. Weak‑tie social connection in turn exhibits a significant positive association with subjective loneliness. The mediation test confirms a statistically significant indirect effect, which means weak‑tie social connection partially mediates the relationship between algorithmic exposure and loneliness. These findings support the connectivity paradox theory: algorithmic recommendation systems can rapidly expand users’ network of casual online acquaintances, yet these weak relationships cannot satisfy the human need for deep, emotionally supportive social bonds, which ultimately elevates subjective loneliness. This study contributes to media sociology by distinguishing algorithmic exposure from general social media consumption, and adds empirical evidence for the “lonely algorithm” framework. Implications for digital literacy education and platform governance are discussed, alongside nuanced reflection on the unintended psychosocial consequences of automated content recommendation systems.

Keywords: algorithmic content exposure; weak‑tie social connection; loneliness; Generation Z; social media; media sociology

Full research content, tables, references and complete data analysis are contained in the attached full-text PDF file.