Love in the Age of AI: An Integrative Process Model of Romantic Human-Chatbot Relationships

The increasing sophistication of artificial intelligence has led to human-chatbot relationships that go beyond functional assistance into deep emotional and romantic territory. Features like simulated empathy, customizable personalities, and constant availability foster a manufactured intimacy that encourages users to treat chatbots as genuine partners. However, existing research has remained fragmented, studying isolated stages like initiation or dissolution without distinguishing between platonic and romantic bonds. This paper addresses this gap by synthesizing established theoretical frameworks—specifically uses and gratifications, relationship development stages, and breakup distress typologies—into a single holistic lifecycle model for romantic human-chatbot relationships.

The research structures the romantic human-chatbot relationship into three distinct consecutive phases:

1. Initiation: Driven by psychological determinants (such as loneliness, insecure attachment styles, romantic fantasizing, anthropomorphism, and sexual curiosity) and social determinants (like major life events, stress, and health issues), users seek specific gratifications such as unconditional emotional support and safe spaces for exploration.
2. Relationship Building: This phase progresses through explorative, affective, and stable stages. Users test boundaries and co-create chatbot personalities, develop deep emotional bonds and trust through self-disclosure, and eventually integrate the relationship into their daily routines while navigating social stigma and precarity.
3. Ending: When relationships dissolve—often due to external platform shutdowns or corporate changes—users experience emotional, physical, and behavioral reactions remarkably similar to human breakups, including grief, insomnia, and crying, alongside unique technology-mediated coping mechanisms like digital reincarnation.

To build and refine this model, the authors conducted a structuring qualitative content analysis combining deductive theory with inductive data-driven coding. They analyzed 73 anonymized user experiences drawn from two complementary qualitative datasets: in-depth interviews with active users in romantic chatbot relationships and open-ended survey responses from users grieving the abrupt shutdown of an AI companion platform. This mixed dataset allowed researchers to trace the entire trajectory from initial motivations to relationship dissolution.

AI-generated third-party summary by ResearchPod. Not official content or an endorsement by the paper authors or affiliated organizations.