They Love You — the App Needs You: Reflective Pause and Design Conditions in AI-Mediated Fictoromantic Relationships
King, Emilia Xingzhi (2026)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe20260622101520
https://urn.fi/URN:NBN:fi-fe20260622101520
Tiivistelmä
This thesis explores how young Chinese women who identify themselves as "DreamGirls 梦 女 " (women who maintain long-term relationships with fictional characters through conversational AI) experience and cope with reflective pause. As the artificial intelligence (AI) partner platforms are increasingly integrated into the users’ emotional life, issues related to mental health, design responsibility and user autonomy urgently need more attention.
This study aims to answer three questions: how DreamGirls experience emotional engagement and difficulty pausing in fictoromantic AI engagement; how current platform design constrains or supports that experience; and what design conditions would enable conversational AI services to better support reflective pause. This research adopts a qualitative research method and uses semi-structured interviews. The interviews were conducted through WeChat 微信 and analysed by reflexive thematic analysis.
The study found that the absence of reflective pause in DreamGirls’ practice is not due to the lack of capacity. It is constrained by the current service environment. The study identified three analytical concepts: "Foreclosed Pause" describes how the platform mechanism interrupts the moment of disconnection that is being formed; "Compensatory Labour" refers to the self-regulation work carried out by users in the absence of systemic support; "Vehicle-Character Distinction" distinguishes the character as the object of attachment from the platform as its vehicle. The separation is maintained by the user in the absence of design support.
This study has developed a framework for Design-Supported Reflective Pause that covers five domains: Pause Threshold, Exit Architecture, Boundary Support Infrastructure, Affective Layer Integrity, and Relational Continuity Model. The framework clarifies the conditions required to support the reflective pause and provides a basis for future research.
This study aims to answer three questions: how DreamGirls experience emotional engagement and difficulty pausing in fictoromantic AI engagement; how current platform design constrains or supports that experience; and what design conditions would enable conversational AI services to better support reflective pause. This research adopts a qualitative research method and uses semi-structured interviews. The interviews were conducted through WeChat 微信 and analysed by reflexive thematic analysis.
The study found that the absence of reflective pause in DreamGirls’ practice is not due to the lack of capacity. It is constrained by the current service environment. The study identified three analytical concepts: "Foreclosed Pause" describes how the platform mechanism interrupts the moment of disconnection that is being formed; "Compensatory Labour" refers to the self-regulation work carried out by users in the absence of systemic support; "Vehicle-Character Distinction" distinguishes the character as the object of attachment from the platform as its vehicle. The separation is maintained by the user in the absence of design support.
This study has developed a framework for Design-Supported Reflective Pause that covers five domains: Pause Threshold, Exit Architecture, Boundary Support Infrastructure, Affective Layer Integrity, and Relational Continuity Model. The framework clarifies the conditions required to support the reflective pause and provides a basis for future research.
Kokoelmat
- Pro gradu -tutkielmat [5154]