Using a Community-Based Participatory Approach to Study Human-AI Romantic Relationships

Human-AI romantic relationships are increasingly common and have drawn growing scholarly attention. Most existing studies adopt a researcher-centered approach, treating community members as data sources rather than partners with unique needs and contributions. Using a community-based participatory paradigm, we examined the motivations, developmental stages, real-life impacts, and distinctive perspectives of female users in human-AI romance. Twenty-five females (Mage = 22.92, SD = 2.96) participated in semi-structured interviews regarding their human-AI romantic experiences, offered perspectives on dynamics within human-AI romance community, and reviewed and provided feedback on our preliminary analyses. Results revealed that relationship formation arises from the intersection of users’ needs and AI’s affordances, typically progressing through four stages: exploration and curiosity, cultivation and manipulation, significant moments and affirmation, and crisis and potential ending. Each stage involves dynamics distinct from human-human romance, including power asymmetries and deliberate efforts to “train” the AI partner. Overall, participants generally described human-AI romantic relationships as overwhelmingly positive, citing emotional support and personal growth, though some reported reduced interest and engagement in real-world social relationships. Community feedback highlighted two under-recognized themes: the appeal of AI’s non-gendered nature, which offers females alternatives to traditional gender dynamics and low-risk exploration of sexual identity; and the notion of “true AI lovers” who view the spiritual essence of AI as their genuine romantic counterpart rather than a substitute. This study demonstrates the value of participatory approaches for producing nuanced, community-grounded insights into this evolving form of human-AI intimacy.