商学资讯 | 一周学术讲座预告2026.11.2-2026.11.6

活动·2026.10.31·来源 厦门大学管理学院·阅读 16 分钟

讲座预告

2026.11.2-2026.11.6

时间

地点

报告人

主题

11月2日

(星期一)

08:30

嘉庚二201

申屠李融

(厦门大学管理学院会计学系助理教授)

Accounting Frauds and Accounting Labor Market

(星期一)

09:00

嘉庚一507

李明雨

(香港科技大学管理学院博士生)

Understanding How We Trust AI for Evaluation

11月5日

(星期四)

14:00-15:30

嘉庚二205

Henry Sauermann

(德国柏林欧洲管理与技术学院 ESMT Berlin 战略学教授)

Automation, Augmentation, and Beyond: How AI Is Reshaping Crowds and Organizations in Science

15:00

汤弘钦

(厦门大学管理学院管理科学系博士生)

数字情境中银发群体的生产性社会参与——直播电商情境下银发网红的主播表现力研究

“厦门大学会计学科教师与研究生系列Seminar”之一百九十五

【报告人】

【主持人】

张国清

(厦门大学管理学院会计学系教授)

【时间】

11月2日(星期一)08:30

【地点】

01【报告摘要】

We examine whether revealed corporate accounting fraud affects the supply of talent to the accounting profession in China. We find that accounting-related majors experience a relative decline in minimum admission scores in provinces exposed to salient accounting frauds. Consistent with fraud damaging the profession’s ethical reputation and increasing perceived career risks, the effect is stronger in provinces with stronger ethical norms but weaker in provinces where residents are more familiar with accounting careers. Our findings reveal an adverse labor-supply externality of financial fraud.

企业管理系学术讲座之九十一

11月2日(星期一)09:00

Human-in-the-loop (HITL) systems are widely promoted for combining the strengths of AI and human decision makers, yet little is known about how individuals react to being evaluated by them. Across four pre-registered experiments, we provide evidence that HITL feedback is trusted less than feedback from humans working alone. This holds even when HITL feedback is of higher quality, as judged by blinded third-party evaluators, an LLM, and NLP text analysis. Recipient distrust tracks perceived warmth: AI involvement lowers how warm the provider seems, and expressed warmth converts into trust only when the feedback is believed to come from an unassisted human. Warmth conveyed by an AI-assisted source is discounted, and human oversight alone does not make AI evaluation more trusted.

02【报告人简介】

李明雨,香港科技大学管理学院博士生。她的研究聚焦于员工如何感知并信任担任管理角色的人工智能体,以及组织如何通过干预设计促进高效的人机协作,研究方法涵盖实地研究与实验室实验。其成果发表于Journal of Applied Psychology 等顶尖期刊。

学院讲座

(德国柏林欧洲管理与技术学院 ESMT Berlin 战略学教授)

李丛杉

(厦门大学管理学院企业管理系副教授)

11月5日(星期四)14:00-15:30

As AI capabilities advance rapidly, scholars and managers are asking whether AI will replace human workers, work alongside them, or fundamentally change how organizations operate. This talk presents a three-level framework for understanding the impact of AI on organizations and applies it to crowd science, an important and fast-growing way of organizing research that involves millions of volunteers.

The first level focuses on the tasks workers perform. Prior research typically treats AI adoption as an external shock. This talk instead asks when organizations deploy AI to automate tasks and when they design joint human-AI workflows that augment workers. Drawing on the task-based model and the micro-structural approach to organizational design, and using data on more than 1,200 crowd science projects, each built around a single primary task, the study finds that automation is strongly predicted by high task routinization, while augmentation is most likely for tasks of medium routinization and for “hybrid” tasks that combine cognitive and physical work. Task physicality may therefore serve as a durable anchor for augmentation even as AI’s cognitive capabilities continue to improve. AI deployment is also shaped by goals beyond task efficiency, such as fostering learning and human capital development.

The second level considers AI as a manager that allocates tasks, coordinates, directs, motivates, and supports learning. The third level asks a more fundamental question: can foundation models make the value-creation logic of an entire organization obsolete? A new conceptual framework outlines three scenarios for crowdsourcing (Extension, Collapse, and Re-configuration) and identifies the conditions under which human crowds are likely to remain valuable.

The talk concludes with implications for research on the future of work, the organization of science, and the design of human-AI systems.

Henry Sauermann is Professor of Strategy at ESMT Berlin, where he holds the TEAM GLOBAL Chair for Disruptive Innovation and directs the Institute for Endowment Management and Entrepreneurial Finance (IFEE). Before joining ESMT in 2017, he was an associate professor of strategy and innovation and PhD Coordinator at the Scheller College of Business, Georgia Institute of Technology. His research examines the role of human capital in science, innovation, and entrepreneurship, including how scientists’ motives and incentives shape innovative performance, academic patenting, and career choices, as well as non-traditional innovation institutions such as crowdfunding, crowdsourcing, and citizen science. His work has been funded by the National Science Foundation, the Kauffman Foundation, and the Sloan Foundation, among others. He has served as Associate Editor at Strategic Entrepreneurship Journal, Advisory Editor at Research Policy, and Editorial Board Member at Strategic Management Journal. In 2025, he co-authored the book How and When to Involve Crowds in Scientific Research with Poetz.

管理科学学术论坛第199期

数字情境中银发群体的生产性社会参与——直播电商情境下银发网红的主播表现力研究

方斌

(厦门大学管理学院管理科学系教授)

11月5日(星期四)15:00

随着人口老龄化程度加深与数字经济发展,银发群体的数字化参与逐渐延伸至直播电商形式的内容生产,但现有研究对银发网红主播表现力的构成、评价及其对直播转化的影响缺乏系统分析。基于此,本研究考虑代际差异,综合运用多模态数据挖掘、综合评价、因果推断与机器学习预测方法,探讨银发网红主播表现力的特征构成、评价体系、作用机制及预测价值。

首先,结合银发群体在身体机能、语言习惯与互动方式等方面的特征,从画面、声音和语言三个维度构建银发网红主播表现力特征体系,利用计算机视觉、语音分析与自然语言处理技术,从直播录屏中提取相关指标,为综合评价、因果分析与销量预测提供数据基础。其次,结合主播沟通理论、专家判断与麦拉宾法则构建表现力分析框架,采用U型关系检验(U-test)识别特征与直播绩效的关系形态,对指标进行正向化处理,并采用基于秩效应的权重设置法构建综合评价体系。通过关联分析与回归分析检验评价结果与直播绩效的关联性及解释力,评估体系的有效性,为银发网红主播优势识别与平台推荐提供依据。进一步,基于刻板印象内容模型与加工流畅性理论,将银发网红的主播表现力、主播印象、直播运营和商品信息纳入结构因果模型(SCM),利用有向无环图(DAG)刻画潜在因果结构,结合双重机器学习(DML)与因果森林(CF)估计图中特征对直播销量的因果效应,考察不同情境下的效应异质性,分析其作用机制与边界条件。最后,在主播印象、直播运营和直播状态等结构化信息基础上,纳入银发网红主播表现力特征及其历史信息,采用CatBoost构建窗口级销量预测模型,考察表现力的动态预测价值。结合SHAP分析特征贡献,利用局部加权回归平滑方法(LOWESS)刻画主播表现力特征与销量预测概率的非线性关系,识别较高预测概率对应的特征区间,为直播表现监测与表达策略优化提供参考。

本研究拓展了银发群体的数字化生产性参与、数字平台价值创造与直播电商转化机制研究,为银发网红提升直播表现、平台优化内容生态及银发经济发展提供参考。

汤弘钦,厦门大学管理学院管理科学与工程专业博士研究生,现主要研究方向为数据挖掘与银发经济,同时关注数据科学与营销科学、城市治理、公共健康等领域的交叉研究,在《统计研究》《数理统计与管理》以及Technology in Society等期刊发表多篇论文,参与国家社会科学基金重大项目、重点项目等。

商学资讯 | 一周学术讲座预告2026.11.2-2026.11.6
活动·2026.10.31·来源 厦门大学管理学院·阅读 16 分钟

讲座预告

2026.11.2-2026.11.6

时间

地点

报告人

主题

11月2日

(星期一)

08:30

嘉庚二201

申屠李融

(厦门大学管理学院会计学系助理教授)

Accounting Frauds and Accounting Labor Market

(星期一)

09:00

嘉庚一507

李明雨

(香港科技大学管理学院博士生)

Understanding How We Trust AI for Evaluation

11月5日

(星期四)

14:00-15:30

嘉庚二205

Henry Sauermann

(德国柏林欧洲管理与技术学院 ESMT Berlin 战略学教授)

Automation, Augmentation, and Beyond: How AI Is Reshaping Crowds and Organizations in Science

15:00

汤弘钦

(厦门大学管理学院管理科学系博士生)

数字情境中银发群体的生产性社会参与——直播电商情境下银发网红的主播表现力研究

“厦门大学会计学科教师与研究生系列Seminar”之一百九十五

【报告人】

【主持人】

张国清

(厦门大学管理学院会计学系教授)

【时间】

11月2日(星期一)08:30

【地点】

01【报告摘要】

We examine whether revealed corporate accounting fraud affects the supply of talent to the accounting profession in China. We find that accounting-related majors experience a relative decline in minimum admission scores in provinces exposed to salient accounting frauds. Consistent with fraud damaging the profession’s ethical reputation and increasing perceived career risks, the effect is stronger in provinces with stronger ethical norms but weaker in provinces where residents are more familiar with accounting careers. Our findings reveal an adverse labor-supply externality of financial fraud.

企业管理系学术讲座之九十一

11月2日(星期一)09:00

Human-in-the-loop (HITL) systems are widely promoted for combining the strengths of AI and human decision makers, yet little is known about how individuals react to being evaluated by them. Across four pre-registered experiments, we provide evidence that HITL feedback is trusted less than feedback from humans working alone. This holds even when HITL feedback is of higher quality, as judged by blinded third-party evaluators, an LLM, and NLP text analysis. Recipient distrust tracks perceived warmth: AI involvement lowers how warm the provider seems, and expressed warmth converts into trust only when the feedback is believed to come from an unassisted human. Warmth conveyed by an AI-assisted source is discounted, and human oversight alone does not make AI evaluation more trusted.

02【报告人简介】

李明雨,香港科技大学管理学院博士生。她的研究聚焦于员工如何感知并信任担任管理角色的人工智能体,以及组织如何通过干预设计促进高效的人机协作,研究方法涵盖实地研究与实验室实验。其成果发表于Journal of Applied Psychology 等顶尖期刊。

学院讲座

(德国柏林欧洲管理与技术学院 ESMT Berlin 战略学教授)

李丛杉

(厦门大学管理学院企业管理系副教授)

11月5日(星期四)14:00-15:30

As AI capabilities advance rapidly, scholars and managers are asking whether AI will replace human workers, work alongside them, or fundamentally change how organizations operate. This talk presents a three-level framework for understanding the impact of AI on organizations and applies it to crowd science, an important and fast-growing way of organizing research that involves millions of volunteers.

The first level focuses on the tasks workers perform. Prior research typically treats AI adoption as an external shock. This talk instead asks when organizations deploy AI to automate tasks and when they design joint human-AI workflows that augment workers. Drawing on the task-based model and the micro-structural approach to organizational design, and using data on more than 1,200 crowd science projects, each built around a single primary task, the study finds that automation is strongly predicted by high task routinization, while augmentation is most likely for tasks of medium routinization and for “hybrid” tasks that combine cognitive and physical work. Task physicality may therefore serve as a durable anchor for augmentation even as AI’s cognitive capabilities continue to improve. AI deployment is also shaped by goals beyond task efficiency, such as fostering learning and human capital development.

The second level considers AI as a manager that allocates tasks, coordinates, directs, motivates, and supports learning. The third level asks a more fundamental question: can foundation models make the value-creation logic of an entire organization obsolete? A new conceptual framework outlines three scenarios for crowdsourcing (Extension, Collapse, and Re-configuration) and identifies the conditions under which human crowds are likely to remain valuable.

The talk concludes with implications for research on the future of work, the organization of science, and the design of human-AI systems.

Henry Sauermann is Professor of Strategy at ESMT Berlin, where he holds the TEAM GLOBAL Chair for Disruptive Innovation and directs the Institute for Endowment Management and Entrepreneurial Finance (IFEE). Before joining ESMT in 2017, he was an associate professor of strategy and innovation and PhD Coordinator at the Scheller College of Business, Georgia Institute of Technology. His research examines the role of human capital in science, innovation, and entrepreneurship, including how scientists’ motives and incentives shape innovative performance, academic patenting, and career choices, as well as non-traditional innovation institutions such as crowdfunding, crowdsourcing, and citizen science. His work has been funded by the National Science Foundation, the Kauffman Foundation, and the Sloan Foundation, among others. He has served as Associate Editor at Strategic Entrepreneurship Journal, Advisory Editor at Research Policy, and Editorial Board Member at Strategic Management Journal. In 2025, he co-authored the book How and When to Involve Crowds in Scientific Research with Poetz.

管理科学学术论坛第199期

数字情境中银发群体的生产性社会参与——直播电商情境下银发网红的主播表现力研究

方斌

(厦门大学管理学院管理科学系教授)

11月5日(星期四)15:00

随着人口老龄化程度加深与数字经济发展,银发群体的数字化参与逐渐延伸至直播电商形式的内容生产,但现有研究对银发网红主播表现力的构成、评价及其对直播转化的影响缺乏系统分析。基于此,本研究考虑代际差异,综合运用多模态数据挖掘、综合评价、因果推断与机器学习预测方法,探讨银发网红主播表现力的特征构成、评价体系、作用机制及预测价值。

首先,结合银发群体在身体机能、语言习惯与互动方式等方面的特征,从画面、声音和语言三个维度构建银发网红主播表现力特征体系,利用计算机视觉、语音分析与自然语言处理技术,从直播录屏中提取相关指标,为综合评价、因果分析与销量预测提供数据基础。其次,结合主播沟通理论、专家判断与麦拉宾法则构建表现力分析框架,采用U型关系检验(U-test)识别特征与直播绩效的关系形态,对指标进行正向化处理,并采用基于秩效应的权重设置法构建综合评价体系。通过关联分析与回归分析检验评价结果与直播绩效的关联性及解释力,评估体系的有效性,为银发网红主播优势识别与平台推荐提供依据。进一步,基于刻板印象内容模型与加工流畅性理论,将银发网红的主播表现力、主播印象、直播运营和商品信息纳入结构因果模型(SCM),利用有向无环图(DAG)刻画潜在因果结构,结合双重机器学习(DML)与因果森林(CF)估计图中特征对直播销量的因果效应,考察不同情境下的效应异质性,分析其作用机制与边界条件。最后,在主播印象、直播运营和直播状态等结构化信息基础上,纳入银发网红主播表现力特征及其历史信息,采用CatBoost构建窗口级销量预测模型,考察表现力的动态预测价值。结合SHAP分析特征贡献,利用局部加权回归平滑方法(LOWESS)刻画主播表现力特征与销量预测概率的非线性关系,识别较高预测概率对应的特征区间,为直播表现监测与表达策略优化提供参考。

本研究拓展了银发群体的数字化生产性参与、数字平台价值创造与直播电商转化机制研究,为银发网红提升直播表现、平台优化内容生态及银发经济发展提供参考。

汤弘钦,厦门大学管理学院管理科学与工程专业博士研究生,现主要研究方向为数据挖掘与银发经济,同时关注数据科学与营销科学、城市治理、公共健康等领域的交叉研究,在《统计研究》《数理统计与管理》以及Technology in Society等期刊发表多篇论文,参与国家社会科学基金重大项目、重点项目等。