Parasocial relationships with artificial intelligence (AI): A systematic review of benefits and risks

Abstract

Advances in artificial intelligence (AI) have transformed AI systems into increasingly interactive and relational agents, raising questions about the consequences of the perceived sense of intimacy and reciprocity formed with AI systems. The present systematic review synthesised empirical evidence on the benefits and risks of AI parasocial relationships across diverse AI agents and contexts. Following PRISMA guidelines, 39 empirical records were identified and analysed using narrative thematic analysis. Through this, five key benefits of AI parasocial relationships were identified, namely personal development and adaptive functioning, emotional support, social needs fulfilment, enjoyment and entertainment, and community participation and social belonging. However, these perceived relationships also carry risks, consisting of commercial persuasion and consumer influence, displacement of human relationships, emotional dependence, privacy and data exploitation concerns, and behavioural addiction and compulsivity. The review also highlights research gaps including the lack of consistent definition and measurement of AI parasocial relationships, and underscores the need for greater theoretical clarity, longitudinal research, and responsible AI design to maximise benefits while mitigating potential harms for users.