时间：2025年12月2日，下午15:30
地点：62#326会议室
组织单位：物流与电子商务学院
主讲人：黄鹂强 教授
演讲内容简介
This study explores the impact of Large Language Model (LLM) assistance on employee task performance in unstructured, higher-level problem-solving contexts, with a focus on the roles of explicit and tacit knowledge integration. Through two field experiments conducted at a Chinese telemarketing firm, we investigate how collaborative work with the LLM, initially fine-tuned with explicit organizational knowledge, and then tuned with tacit input from front-line telemarketers, affects sales personnel performance. The first experiment reveals a skill-biased effect: while low and middle performers benefit significantly from LLM guidance, top performers experience a decline in performance. Addressing this gap, the second experiment introduces an annotation feature enabling top performers to contribute their situational responses, thereby integrating tacit knowledge into the LLM. Results show that this enhancement mitigates the negative effect on top performers and further boosts performance across all skill levels. Mediation analyses identify convergence towards LLM recommendations, through enhanced standardization and reduced policy violations, as the key mechanism driving performance improvements. Our findings contribute to the literature by positioning LLMs as knowledge repositories, converting tacit knowledge that was unconvertable before into explicit and actionable knowledge that is ready for application, and offering practical insights for organizations to optimize AI-augmented workflows while preserving human expertise.
主讲人简介
黄鹂强博士现任浙江大学教授、博士生导师。黄教授以信息系统战略与战略信息系统，数据要素与数据资产管理为主要研究领域。他的研究受到了科技部国家重点研发计划项目、国家社会科学基金重大项目、国家自然科学基金重点项目以及浙江省杰出青年基金和众多企业的资助，研究成果发表于国内外学术期刊如MIS Quarterly, Information Systems Research,《管理工程学报》等，目前担任国际学术期刊Journal of Global Information Management和Information Technology & Management副主编。此外，黄教授的研究成果也受到了多家知名企业应用，相关资政报告受到了省部级领导的批示。
欢迎感兴趣的老师、同学参加！