联系我们

论文查询结果	返回搜索

论文编号：	16011
作者编号：	2120243856
上传时间：	2026/6/2 18:49:51
中文题目：	对话式AI的赞美具体性对用户—AI关系质量的影响研究
英文题目：	A Study on the Impact of Praise Specificity in Conversational AI on User–AI Relationship Quality
指导老师：	任星耀
中文关键字：	对话式AI；赞美具体性；用户—AI关系质量；心智感知；AI角色；任务复杂性
英文关键字：	Conversational AI; Praise Specificity; User-AI Relationship Quality; Mind Perception; AI Role; Task Complexity
中文摘要：	随着人工智能与自然语言处理技术的快速发展，对话式AI已从传统功能性工具逐步转型为具备社交属性、能够开展情感交互的智能主体，广泛融入办公协作、学习教育、智能客服等多元场景。言语线索作为人机沟通的核心载体，其反馈方式在塑造用户—AI关系质量中发挥着日益重要的作用。赞美作为人际互动中普遍且高频的言语行为，具有传递关注、表达认可、增进情感联结的功能。然而在人机交互情境下，对话式AI的赞美如何影响用户—AI关系质量，其内在机制与边界条件仍有待探讨。 本研究基于计算机作为社会行动者理论、心智感知理论和社会角色理论，构建包含主效应、中介效应与调节效应的研究模型，探究对话式AI赞美具体性对用户—AI关系质量的影响机制及边界条件。研究采用实验法，通过预实验筛选实验操控材料后开展三个正式实验：实验一采用单因素两水平（赞美具体性：高 vs. 低）的组间设计，验证赞美具体性的主效应及心智感知的中介作用；实验二和实验三则分别采用2（赞美具体性：高 vs. 低）×2（AI角色定位：伙伴 vs. 工具）和2（赞美具体性：高 vs. 低）×2（任务复杂性：高 vs. 低）组间设计，检验AI角色定位和任务复杂性的调节作用。 研究结果表明：（1）赞美具体性对用户—AI关系质量具有正向影响，高具体性赞美能更有效地提升感知交互质量、满意度、信任及持续使用意愿；（2）心智感知在赞美具体性与关系质量之间发挥中介作用；（3）AI角色定位是重要边界条件，伙伴定位下高具体性赞美效应更突出，工具定位下差异不显著；（4）任务复杂性构成另一边界条件，复杂任务中主效应显著，简单任务中不显著。 在理论层面，本研究将赞美具体性作为对话式AI言语线索研究的新变量，关注其对用户—AI关系质量的影响机制与作用边界。在实践层面，研究为对话式AI的交互设计、差异化角色定位策略及基于任务情境的动态反馈调整提供了科学依据，有助于提升智能客服、虚拟助手等的用户体验与关系维系能力。
英文摘要：	With the rapid advancement of artificial intelligence and natural language processing technologies, conversational AI has gradually transformed from traditional functional tools into intelligent agents with social attributes and the capacity for emotional interaction, becoming widely integrated into diverse scenarios such as office collaboration, educational learning, and intelligent customer service. As the core carrier of human-computer communication, feedback patterns of verbal cues play an increasingly important role in shaping the quality of user-AI relationships. Praise, as a common and frequent verbal act in interpersonal interaction, serves to convey attention, express recognition, and enhance emotional connection. However, in human-computer interaction contexts, how praise from conversational AI affects user-AI relationship quality, along with its underlying mechanisms and boundary conditions, remains to be explored. Grounded in the Computers Are Social Actors theory, Mind Perception theory, and Social Role theory, this study constructs a research model encompassing main effects, mediating effects, and moderating effects to investigate the mechanisms and boundary conditions through which conversational AI praise specificity influences user-AI relationship quality. Adopting an experimental approach, this study conducted three formal experiments after screening manipulation materials via a pre-experiment. Experiment 1 adopted a one-factor, two-level (praise specificity: high vs. low) between-subjects design to verify the main effect of praise specificity and the mediating effect of mind perception. Experiment 2 and Experiment 3 adopted 2 (praise specificity: high vs. low) × 2 (AI role positioning: partner vs. tool) and 2 (praise specificity: high vs. low) × 2 (task complexity: high vs. low) between-subjects designs respectively to test the moderating effects of AI role positioning and task complexity. The findings show that: (1) praise specificity positively influences user-AI relationship quality, with highly specific praise more effectively enhancing perceived interaction quality, satisfaction, trust, and continuance intention; (2) mind perception mediates the relationship between praise specificity and relationship quality; (3) AI role positioning serves as an important boundary condition, with the effect of highly specific praise being more pronounced under the partner positioning and non-significant under the tool positioning; (4) task complexity constitutes another boundary condition, where the main effect is significant in complex tasks but not in simple tasks. Theoretically, this thesis introduces praise specificity as a new variable into the research on linguistic cues in conversational AI, and reveals its influencing mechanism and boundary conditions on user–AI relationship quality. Practically, it provides a scientific basis for the interaction design of conversational AI, differentiated role positioning strategies, and dynamic feedback adjustment based on task scenarios, helping to improve the user experience and relationship maintenance capability of systems such as intelligent customer service and virtual assistants.
查看全文：	预览  下载(下载需要进行登录)