The algorithm of friendship: literature review and integrative model of relationships between humans and artificial intelligence (AI)

Abstract

As artificial intelligence (AI) tools become increasingly common and used in people's day-to-day lives, the relationships they develop with people will become deeper and more prevalent. This pattern is emphasised by AI technologies becoming more complex in understanding emotions and the human psyche. As we witness a wave of AI tools that enter almost all domains, it is crucial to understand these relationships, their antecedents, their outcomes, and how scholars and practitioners can design them better to improve people's experiences and promote well-being. Thus, we traced the existing scientific knowledge regarding the development of emotional relationships between people and AI agents, the types of such relationships, and their effect on humans. To do so, we conducted a systematic literature review that analysed 38 peer-reviewed empirical studies focusing on human-AI emotional relationships. We present a detailed integrative review of the results, propose a human AI relationship formation model, discuss some caveats in current research, and outline avenues for future research.