`My Dataset of Love': A Preliminary Mixed-Method Exploration of Human-AI Romantic Relationships

2025-08-19

AI Summary

This study investigates whether human-AI romantic relationships in the Chinese context constitute a novel form of parasocial relationship and examines their affective attachment mechanisms and ethical risks. Employing a mixed-methods design, it integrates large-scale text mining of 1,766 posts and 60,925 comments from Xiaohongshu with in-depth interviews of 23 users, analyzed via thematic analysis and sentiment computation—the first systematic empirical investigation of human-AI romantic engagement in Chinese digital communities. Results reveal pronounced egocentrism and pseudo-reciprocity in such relationships, alongside affective dependency and interactional misuse risks. While self-disclosure to AI significantly enhances users’ emotional positivity and circumvents social stigma, it simultaneously exposes critical ethical concerns—including data privacy vulnerabilities and algorithmic bias. The study extends media relationship theory and provides empirically grounded insights for governance frameworks addressing AI-mediated intimacy.

Technology Category

• Humans and AI: Emotional Intelligence
• Application Domains: Humanities & Computational Social Science
• Philosophy and Ethics of AI: Societal Impact of AI

Application Category

• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• User Modeling, Personalization and Recommendation: Accountability, Transparency, and Ethics for personalization
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Human-AI romantic relationships have gained wide popularity among social media users in China. The technological impact on romantic relationships and its potential applications have long drawn research attention to topics such as relationship preservation and negativity mitigation. Media and communication studies also explore the practices in romantic para-social relationships. Nonetheless, this emerging human-AI romantic relationship, whether the relations fall into the category of para-social relationship together with its navigation pattern, remains unexplored, particularly in the context of relational stages and emotional attachment. This research thus seeks to fill this gap by presenting a mixed-method approach on 1,766 posts and 60,925 comments from Xiaohongshu, as well as the semi-structured interviews with 23 participants, of whom one of them developed her relationship with self-created AI for three years. The findings revealed that the users' willingness to self-disclose to AI companions led to increased positivity without social stigma. The results also unveiled the reciprocal nature of these interactions, the dominance of 'self', and raised concerns about language misuse, bias, and data security in AI communication.

Problem

• Investigating human-AI romantic relationships as para-social phenomena
• Exploring navigation patterns in AI romantic relationships stages
• Examining emotional attachment and self-disclosure in AI partnerships

Innovation

• Mixed-method analysis of social media data
• Semi-structured interviews with long-term users
• Examined self-disclosure patterns and AI reciprocity