Human or AI Fortune-Telling: Exploring Users’ Expectations and Trust Formation in Relationship Advice-Seeking with Human vs. AI I-Ching Divination

CSCW has long investigated computational support for social connection and intimate relationships, while people have also used divination for inquiring into romantic relationships for centuries. This paper investigates computer-mediated romantic-relationship advice-seeking with I-Ching, an ancient Chinese divination system. Participants provided a romantic relationship query and received two textual responses, one from a human I-Ching fortune-teller and the other from an LLM I-Ching tool. Participants were asked to identify which response was human-generated and interpret the responses to aid their romantic situation. Our findings reveal content-based cues with which participants distinguished the response source (a human fortune-teller or LLM); and further articulate how specific response features shaped their perceptions of credibility. We discuss how trust and credibility in LLM-mediated romantic advice-seeking are shaped by affective alignment and interpretive openness. Drawing on divinatory thinking and swift trust, we call for relational, experience-centered frameworks for evaluating LLMs in intimate, emotionally complex contexts.