嘉宾简介：陈亚盛，博士、教授、博士生导师、加拿大注册会计师（CPA）、英国皇家特许管理会计师（ACMA）、全球特许管理会计师（CGMA），获加拿大西安大略大学毅伟商学院博士学位，曾先后在加拿大西安大略大学和西蒙菲莎大学任教 10 年。研究领域为管理会计理论、神经会计研究方法、管理控制系统设计、人工智能在会计中的应用。曾主持加拿大国家人文与社会科学基金项目 3 项、加拿大国家会计学会研究基金项目 1 项，国家自然科学基金面上项目2项、福建省高校领军人才资助计划，参与教育部人文社会科学重点研究基地重大项目、国家财政部管理会计专项课题研究项目。在 Journal of Accounting Research、 Accounting and Finance、 World Economy、 Current Psychology、Sustainability、Journal of International Accounting Research等国际会计顶尖学术期刊和国内外知名学术期刊上发表多篇论文。入选福建省和厦门市高层次引进人才，厦大南强青年拔尖人才，澳门新葡平台网址8883网站群贤计划。目前从事研究课题包括将眼动跟踪和脑扫描等神经会计学研究方法与人工智能算法相结合，设计促进企业创新的智能管理控制系统。

论文摘要：Companies are increasingly applying artificial intelligence (AI) technology to employee performance management, including automated negotiation and performance target setting. However, existing research has not fully explored the underlying mechanism through which AI supervisors influence employees behavior and performance. To compare the effect of AI supervisors and human supervisors on the behavior and performance of employees in target setting negotiations, we conducted a 2 x 2 experiment using an electrodermal activity device to measure the emotional arousal of participants. The results indicate that compared to AI supervisors, human supervisors lead to higher levels of employees' emotional arousal during the negotiation of performance targets, resulting in increased employee engagement in bargaining and improved employee task performance.Adopting a more comprehensive automation-augmentation paradox perspective, we argue that negotiation tactics moderate the relationship between supervisor type and employee emotional arousal. Specifically, when supervisors make concessions later in the negotiation process, employees negotiating with a human supervisor exhibit higher levels of emotional arousal, resulting in lower concessions and higher task performance compared to those negotiating with an AI supervisor. However, when supervisors' concessions occur earlier, there is no significant difference observed regarding their impact on employee negotiation behavior and task performance between the human supervisor groups and AI supervisor groups. In summary, our findings enrich the literature on applying AI in management accounting while also providing important implications for optimizing organizational management controls and improving employee performance in the era of artificial intelligence.