题　目：Are AI the Picky Investors We Need? Scrutinizing Executive Logic with Large Language Models

嘉　宾：杨楷 助理教授 深圳大学

主持人：万相伟 副教授 上海交通大学安泰经济与管理学院

时　间：2026年4月15日（周三）13:30-15:00

地　点：上海交通大学 徐汇校区安泰浩然楼306

内容简介:

We utilize large language models (LLMs) to systematically analyze five types of logical flaws in executive responses during earnings conference calls, specifically irrelevance, inconsistency, insufficiency, inappropriate presumption, and ambiguity. We develop FinWise, a framework employing Chain-of-Thought prompting and reflection mechanisms to quantify and qualitatively explain these flaws. Validating FinWise on U.S. earnings calls from 2005 to 2024, we find that a higher intensity of logical flaws is associated with lower short-window abnormal returns and weaker price responses to earnings news. Additional tests suggest that these logical flaws capture information processing frictions rather than negative fundamental news. Overall, this study demonstrates that appropriately guided LLMs can effectively emulate investor perceptions and generate economically meaningful indicators of communication quality.

演讲人简介：

Kai Yang is an Assistant Professor at the College of Economics, Shenzhen University. He obtained his Ph.D. from the College of Business at the City University of Hong Kong.He serves as a committee member of the Social Media Processing Committee and Affective Computing Committee under the Chinese Information Processing Society of China (CIPSC). He also acts as a reviewer for several top journals in the field, including MIS Quarterly (MISQ), Information Systems Research (ISR), Journal of Commerce (JOC), and Decision Support Systems (DSS).

His research focuses primarily on artificial intelligence, financial technology (FinTech), and natural language processing (NLP). His core research centers on the application of affective computing in business and information management. By integrating multidisciplinary technologies, he constructs intelligent perception models for unstructured data, providing novel approaches for business analysis and decision-making. His work has been published in leading journals such as Information Systems Research (UTD24), Decision Support Systems, and others.

欢迎广大师生参加!