信息管理与商业智能系学术讲座（4月29日）

时间：2025-04-29 13:30-14:30

地点：李达三楼105室

题目：Model Explanation, Accuracy Feedback, and Use of Algorithmic Advice: A Field Experiment

主讲人：曹仔科长聘副教授浙江大学

主持人：李根副教授信息管理与商业智能系

讲座摘要：

The “black-box” nature of modern AI algorithms remains a significant obstacle to building strong human trust in AI systems. By collaborating with a large Chinese retail chain for fresh goods, we conducted a large-scale field experiment, involving 230 brick-and-mortar stores and over 1,000 products (SKUs), to examine how the provision of model explanations and accuracy feedback affects human store managers’ trust in and subsequent use of the algorithmic advice on order quantities. Model explanations are information disclosed to make the underlying logic of the algorithmic advice more transparent, while accuracy feedback is information indicating whether the algorithm is accurate in predicting the actual sales quantities in the past two weeks. We randomly assigned all the stores into four conditions: i) no provision of extra information; ii) providing only model explanations; iii) providing only accuracy feedback; and iv) providing both model explanations and accuracy feedback. The results show that both model explanations and accuracy feedback can help promote trust and adoption, but accuracy feedback is much more influential than model explanations. Moreover, providing both types of transparency information leads to an unexpected negative interaction effect. Our findings bear novel and important implications for AI designers’ trust-building strategies in practice.

主讲人简介:

浙江大学管理学院长聘副教授，博士生导师，获香港科技大学资讯系统 (Information Systems) 博士学位，浙江大学信息管理与信息系统学士学位。研究方向主要为在线用户行为决策分析和在线社区平台治理策略。多篇论文发表于Information Systems Research (UTD24期刊), MIS Quarterly (UTD24期刊)，Journal of Management Information Systems (FT50期刊) 等顶级期刊。博士论文获得2017年度INFORMS信息系统协会Nunamaker-Chen博士论文奖并列第二名 (second runner-up)。研究论文获得2021年度INFORMS电子商务领域 (eBusiness Section) 最佳论文奖第一名 (winner)。主持国家自然科学基金青年项目和优秀青年项目。