谁决定我们的情感生活？——从Moxie机器人“死亡”看情感AI时代的情感不公正

2024年底，情感AI领域发生了一起引发广泛关注的事件——Moxie社交机器人因所属公司停运而全面失效，引发部分用户（尤其是儿童）悲伤与哀悼。尽管当前情感AI的伦理讨论已涵盖隐私操控等议题，然而这种讨论尚未触及情感AI对人类作为“情感存在”所造成的结构性伤害。为进一步拓展情感AI规范性分析的维度，需引入Gallegos提出的“分配式情感不公正”与Krueger所论的“生态式情感不公正”两个理论视角，并结合Moxie案例展开探讨，分析情感AI如何系统性地剥夺个体的情感自由、情感资源与机会及情感承认，甚至导致情感自主性的丧失。在此背景下，情感支架的失效亦可被理解为技术控制与用户自主性之间权力不对等的具体体现。为回应这一问题，认知赋权可为现实中回应情感不公正的权宜应对措施。

At the end of 2024, Moxie, a social robot, drew widespread attention for becoming entirely nonfunctional following the shutdown of its parent company, triggering grief and mourning among its users. So far, the normative dimension of the users' affective experience of Emotional AI, such as in the Moxie case, has received little attention in the field of AI ethics. More specifically, the current ethical discourse has yet to confront the systematic harm Emotional AI could inflict upon human beings as affective beings. In this paper, I shall introduce two approaches to affective injustice, i.e., Gallegos's distributive approach and Krueger's ecological approach, to explore the normative-affective dimension of Emotional AI. Using the Moxie case, I shall demonstrate how Emotional AI can systematically deprive individuals of affective goods, thereby leading to the erosion of one's affective agency, and how an unequal power relationship between technology companies and users in creating their affective environment.