时间：2025-06-18 10:30-12:00

地点：中关村校区主楼四层429

报告人：梁会刚

报告人职务：博士生导师、FedEx MIS 卓越主席

报告人职称：教授

报告人工作单位：孟菲斯大学（University of Memphis）

报告人简介：梁会刚（Huigang Liang）在奥本大学（Auburn University）获得计算机科学硕士学位和医疗保健信息系统博士学位。他的研究兴趣涵盖多种情境下个体和组织层面的社会行为学、管理学及战略性信息技术问题。他的研究成果发表于 MIS Quarterly, Information Systems Research, Journal of MIS, Journal of AIS, MIT Sloan Management Review, Communications of the ACM, Decision Support Systems, Information Systems Journal 和 Journal of Strategic Information Systems 等期刊。他的研究获得了美国国立卫生研究院（NIH）、联邦快递（FedEx）和斯隆基金会（Sloan Foundation）的资助。他曾担任 MIS Quarterly (MISQ) 的副主编，现任 Journal of the AIS (JAIS) 的高级主编，以及 Information Systems Research (ISR) 和 Information & Management 的副主编

报告内容简介：The benefits of generative AI are accompanied by significant risks that demand careful consideration and proactive management. Since the release of ChatGPT at the end of 2022, AI-related incidents that compromise data integrity, authenticity, and confidentiality have been skyrocketing. These emerging threats have garnered tremendous global attention. However, there is no empirical evidence to demonstrate if ChatGPT has contributed to the rise in AI incidents, nor is there clarity on effective strategies for risk mitigation. This concern is especially urgent at the national level, since countries vary in AI development and may experience AI risks differently, and effective AI risk mitigation requires international cooperation. Against this backdrop, this research aims to examine if the advent of generative AI led to a significant increase in AI incidents globally and how national AI investments and institutional quality influenced this effect. We compiled a unique country-level panel dataset from multiple sources, encompassing 10,332 observations from 126 countries worldwide and spanning 82 months, from January 2018 to October 2024. We employ a Difference-in-Differences (DID) strategy to estimate the treatment effect of ChatGPT access on AI incidents. The findings indicate that after the ChatGPT release, countries with ChatGPT access experienced significantly more AI incidents than those without access. Moreover, both AI investment and institutional quality weaken the treatment effect of ChatGPT access on AI incidents. This research contributes to research by establishing the causal link between GAI and AI Incidents while identifying boundary conditions of this effect at the national level. It also contributes to practice by providing actionable insights for policy and governance regarding AI development from a global perspective.

（承办：管理工程系、科研与学术交流中心）