ISSN 1671-3710
CN 11-4766/R
主办:中国科学院心理研究所
出版:科学出版社

• •    

数智时代人机协同绩效反馈的模式及其多层次绩效影响机制

郑立勋, 郑晓明   

  1. 清华大学经济管理学院, 北京 100084 中国
  • 收稿日期:2026-06-30 修回日期:2026-09-15 接受日期:2026-10-22
  • 基金资助:
    国家资助博士后研究人员计划(GZC20250485); 国家自然科学基金项目(72502126); 中国博士后科学基金第78批面上资助(2025M780662); 国家自然科学基金项目(72172074)

Patterns and Multilevel Influence Mechanism of Human–AI Collaborative Performance Feedback in the Digital Intelligence Era

  1. School of Economics and Management, Tsinghua University 100084, China
  • Received:2026-06-30 Revised:2026-09-15 Accepted:2026-10-22
  • Supported by:
    the National Natural Science Foundation of China(72172074)

摘要: 数智技术正在深度嵌入组织绩效管理,绩效反馈逐渐从领导者单独主导转向人工智能(AI)与领导者共同参与的协同反馈。已有AI绩效反馈研究主要围绕AI作为单一反馈来源的配置效应与披露效应展开,较少解释AI与领导者如何在绩效反馈中进行职责配置,以及不同人机配置如何影响团队与员工绩效。围绕上述问题,本研究拟开展三个相互关联的子研究。首先,研究1界定人机协同绩效反馈的内涵,构建“AI自主性(替代/增强)×反馈角色定位(裁判/教练)”的构型框架并开发测量工具。在此基础上,研究2和3进一步聚焦AI提供认知支持与领导进行决策主导的AI增强教练反馈模式。其次,基于工作重塑理论,研究2揭示该模式如何通过团队工作重塑促进团队绩效。最后,基于主动动机模型,研究3探讨该模式如何通过领导反馈效能感、反馈责任感和积极情绪转化为领导发展型反馈行为,进而促进团队和员工绩效。由此,本研究形成“模式建构—模式聚焦—双路径机制解释”的总体框架,将AI绩效反馈研究从反馈来源比较推进到反馈系统内部的人机配置,并从团队集体过程与领导主动转化两个层次揭示其作用机制。

关键词: 人机协同, 绩效反馈, 工作重塑, 主动动机, 工作绩效

Abstract: Digital intelligence technologies are becoming increasingly embedded in organizational performance management, shifting performance feedback from a process primarily led by leaders to one in which artificial intelligence (AI) and leaders assume interlocking responsibilities. Existing research on AI-enabled performance feedback has largely focused on the deployment and disclosure effects of AI as a feedback source, while paying less attention to how AI and leaders allocate and coordinate feedback responsibilities and how different human–AI collaboration ultimately affects team and employee performance. To address the above questions, this research comprises three interrelated studies. First, drawing on research on human–AI collaboration and performance feedback, Study 1 defines human–AI collaborative performance feedback and develops a configurational framework based on AI autonomy (replacement vs. augmentation) × feedback role orientation (judge vs. coach), together with corresponding measures. Building on this framework, Studies 2 and 3 focus on the AI-augmented coaching mode, a structural configuration characterized by AI cognitive support and leader decision authority. Second, drawing on job crafting theory, Study 2 examines how the AI-augmented coaching mode promotes team performance through team job crafting. Third, drawing on the proactive motivation model, Study 3 investigates how this mode enhances leaders’ feedback efficacy, felt responsibility for feedback, and positive affect, thereby promoting leader developmental feedback behavior and ultimately improving team and employee performance. Together, the three studies establish an integrated framework that moves from configuration building to a focused, dual-path explanation of the underlying mechanisms. By doing so, this research advances AI performance feedback research from comparing feedback sources to examining the internal configuration of human–AI feedback systems, while revealing its performance implications through two complementary mechanisms: team-level collective adjustment and leader-driven proactive transformation.

Key words: human–AI collaboration, performance feedback, job crafting, proactive motivation, job performance