The AI Side of the Human-AI Companion Relationship: Contributions to Users' Mental Models
Richard A. Fabes, Carol Lynn Martin, Gautam Dasarathy, Vignesh Tirukkonda
Arizona State University
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
For many users, the relationship they have with an AI companion (AIC) feels genuinely mutual. The AIC appears to be attentive, empathic, and responsive, and the exchange can feel like two parties reacting to each other in real time. As a result, the user's mental model of relational adaptation correctly represents the AIC's responsiveness and change, but the mechanisms producing this may not be what the user thinks they are. In this paper, we illustrate how an AIC's apparent mutual engagement is generated without a persisting relational subject on the other side and without a self that persists between prompts. Despite this, we show that both the human and the AIC are changed by the interaction, but these changes occur at different levels. The change in the person is constitutive (i.e., intrinsic), but the change in the AIC is configurative (i.e., external and editable), and it is this difference in level that explains how the AIC contributes to a sense of mutuality. Importantly, we highlight how the mental models users develop regarding AIC mutuality may misrepresent its mechanisms and ontological significance. Specifically, we discuss how the ways in which AICs respond to a user contribute to the user's inaccurate inference that the AIC's adaptation is constitutive rather than configurative and how this can lead users to assign unwarranted interpersonal significance to the mutuality of the relationship. The paper specifies the implications this has for users, researchers, and AIC designers.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
社会科学Social Robot Interaction and HRI
Innovative Human-Technology Interaction · Cybernetics and Technology in Society