When AI becomes a friend: Emotional support, social displacement, and university belonging among Nigerian freshmen
Main Article Content
Abstract
The growing use of artificial intelligence (AI) companions and virtual assistants among university students raises concerns about whether they alleviate loneliness and enhance belonging or substitute for offline relationships, particularly during the socially demanding transition into university. However, little is known about how freshmen experience this tension in under-resourced higher education contexts such as Nigeria. This study investigated the paradoxical role of AI companions in shaping loneliness, belonging, and emotional well-being among freshmen during this transition. Using a qualitative case study within an interpretivist paradigm, semi-structured interviews and focus group discussions were conducted with 36 participants: 30 freshmen at the University of Uyo, Nigeria, reporting daily relational AI use exceeding 30 minutes, and 6 student affairs counsellors. Five themes emerged: (a) AI companions as emotional scaffolding, (b) the displacement dilemma, (c) the authenticity paradox, (d) institutional belonging in the age of algorithmic companionship, and (e) the conditional role of offline social capital. Critically, even minimal offline social anchors like a roommate, religious fellowship, or study partner appeared to determine whether AI use remained supplementary or became substitutive. This resembled a contextual "digital Matthew effect" where students with stronger offline social resources used AI more complementarily, whereas those with fewer offline anchors described more substitutive use. The study concludes that relational AI can complement or substitute for offline social relationships among university freshmen, depending on their offline social resources and patterns of use. It recommends digital literacy, early identification of substitutive AI use, and strengthened opportunities for meaningful offline connection
Downloads
Article Details

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This article is licensed and distributed under a Creative Common Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA).