AI使用羞耻感：结构维度、测量与创新行为效度检验

AI Usage Shame: Structural Dimensions, Measurement, and Validation of Its Effects on Innovative Behavior

摘要:

在人工智能技术深度融入职场实践的背景下，AI工具凭借其强大的信息处理与创新能力，已成为驱动个体效能提升与组织创新的关键利器。然而，技术赋能的同时却衍生出一种未被充分认识的复杂心理现象“AI使用羞耻感”，即员工因使用AI工具而感知到自身专业能力贬损与行为道德正当性存疑的双重情感体验。现有研究尚存显著局限，概念内涵模糊不清，缺乏本土化测量工具，影响机制不明。鉴于此，研究立足中国组织情境，基于扎根理论方法，通过三级编码提炼出“能力贬损”与“道德焦虑”二维度理论结构，首次系统界定了其概念内涵；据此开发包含10个题项的AI使用羞耻感本土化量表，经验证性因子分析证实所开发测量工具的信效度良好；并进一步整合社会认知理论与资源保存理论，运用本研究开发的AI使用羞耻感量表验证了其对创新行为的理论关联路径；基于537份员工数据发现，AI使用羞耻感与员工创新行为负向关联，自我效能与员工?AI协作能力在二者间起中介作用；人工智能总体态度正向调节员工?AI协作能力的中介路径，而对自我效能的中介路径无显著调节作用。研究通过系统性概念构建、测量工具开发与效度验证，拓展了工作场景中AI使用羞耻感的理论认知，为企业驱动人机协作创新提供了理论依据和管理启示。

Abstract:

Artificial intelligence is deeply integrated into workplace practice. AI tools are regarded as critical drivers for individual efficiency improvement and organizational innovation. A complex psychological phenomenon named AI usage shame is observed, which refers to employees’ dual emotional experience of perceived competence devaluation and questioned moral legitimacy when using AI tools. Significant limitations are found in existing studies, including ambiguous conceptual connotation, lack of localized measurement tools, and unclear influencing mechanisms.Grounded theory is adopted in this study under the Chinese organizational context. A two?dimensional theoretical structure containing competence devaluation and moral anxiety is extracted through three?level coding. The conceptual connotation of AI usage shame is systematically defined for the first time. A localized 10?item scale for AI usage shame is developed. Good reliability and validity of the scale are confirmed by confirmatory factor analysis.Social cognitive theory and conservation of resources theory are integrated. The scale developed in this research is used to test the correlation mechanism between AI usage shame and innovative behavior. Data from 537 employees are analyzed. A negative correlation between AI usage shame and employees’ innovative behavior is verified. Self?efficacy and employee?AI collaboration ability are proven to play mediating roles in the above relationship. Employees’ overall attitudes toward artificial intelligence positively moderate the mediating path of employee?AI collaboration ability, while no significant moderating effect is detected for the mediating path of self?efficacy.This study expands the theoretical understanding of AI usage shame in workplace scenarios through systematic conceptual construction, measurement development and empirical validity test. Theoretical basis and management enlightenment are provided for enterprises to promote human?AI collaborative innovation.