分享人：王晨

地点：经管中楼302

主办方：经济与管理学院

开始时间：2026年05月21日10:00

结束时间：2026年05月21日11:00

学习交流主题：Emotional Expression of Companion AI and Willingness to Use: The Emotions as Social Information (EASI) Perspective

交流内容简介：As companion AI is increasingly deployed in information-overload environments, understanding how AI’s vocal emotional expressions shape users’ willingness to use (WTU) has become critical. Yet research on voice-based emotional communication between LLM-based emotional companion AI (eAI) and users remains lacking. Drawing on EASI theory, we examine whether, under what boundary conditions, and through what pathway mechanisms real-time vocal emotional expressions of eAI causally influence WTU. We fine-tune the Doubao LLM to develop AI exhibiting positive or negative emotional expressions and conduct three experiments. Study 1 establishes the effect of AI emotional valence on WTU under high versus low information load. Study 2 employs deep learning-based causal inference and Qwen2-Audio to validate this bounded causal relationship and test emotional contagion. Study 3 examines social-relational factors. Overall, this research identifies information load as a critical cognitive boundary and emotional contagion as a condition-bounded mechanism for designing emotional companion AI systems.

分享人简介：王晨，福州大学2022级管科专业博士。主要研究领域包括智能管理与决策、管理信息系统；在心理学报（国内顶刊）、工程管理科技前沿（FMS管理科学高质量T1类）、Resources Policy、Neuropsychologia等期刊发表论文，并在Computers in Human Behavior（国际顶刊）、Information Systems Research（UTD24）期刊投稿并under Review两篇论文。