Talking to AI Like a Friend: Emotional Attachment to Artificial Intelligence among Students

Abstract

The swift adoption of AI chatbots in higher education is making these technologies not just functionally useful, but also socially and emotionally meaningful, but the affective implications have been under-theorized. This study investigates the effects of emotional attachment to AI chatbots on motivation to learn languages, psychological well-being, and academic success as well as the cultural moderating factors of such effects. Semi-structured interviews were conducted with 50 students from Bangladesh, China, Germany and the US using an exploratory qualitative design based on the theoretical frameworks of Self-Determination Theory, Social Presence Theory and the CASA paradigm. Reflexive Thematic Analysis found that emotional attachment to AI is a complement rather than a substitute for human social relations. The social presence of AI was found to mediate students’ willingness to communicate in a second language. The collectivist students felt more engaged than the individualists. AI satisfied all three SDT needs, despite autonomous motivation decreasing after eight weeks of exposure. The best affective results were obtained with the RAG-powered AI. The communication with AI peers was most effective to mitigate feelings of shame about language learning in rural and socio-economically deprived students. The findings informed the development of the Integrated Attachment–Motivation–Pedagogy (AMP) model with direct implications for designing AI in culturally sensitive pedagogical practices. It will be interesting to validate the Integrated AMP model in future research through longitudinal studies with structural equation modelling and experimental designs and develop interventions that preserve the motivational and equitable properties of emotional attachment to AI without sacrificing students’ autonomy.