【7月9日】【首都数字经济高质量发展创新中心系列讲座】

The Risk Paradox of Generative AI Diffusion: ChatGPT Exposure, AI Democratization, and AI Investment

Huigang Liang

Generative AI can generate substantial social and economic benefits, but it may also create new forms of AI risk by accelerating the diffusion and development of AI capabilities. Despite widespread concerns about generative AI, rigorous cross-national evidence on its broader risk consequences remains limited. We examine whether country-level access to ChatGPT increased AI risk and identify the mechanisms through which this effect occurs. Using a unique country-month panel covering 131 countries from 2020 to 2025, we employ a difference-in-differences design that compares countries with and without ChatGPT access before and after its public release. We find that countries with ChatGPT access experienced a significant increase in AI risk, measured by non-generative-AI incidents, relative to countries without access. We further reveal two competing mechanisms. ChatGPT exposure increases AI democratization, which amplifies AI risk by expanding decentralized and weakly governed AI development. In contrast, ChatGPT exposure also increases AI investment, which mitigates AI risk by strengthening technical and organizational capacity for safer AI development. In addition, the risk-amplifying effect of AI democratization outweighs the risk-mitigating effect of AI investment. This study contributes to IS research by confirming the link between ChatGPT exposure and country-level AI risk, extending risk society theory to the generative AI era, and identifying competing mechanisms through which frontier AI diffusion produces and contains manufactured risk. It also offers practical implications for policymakers seeking to balance AI democratization, investment, and governance.

讲座时间

2026年7月9日（星期四），9：00-11:00

地点

学院南路校区实验楼303会议室

主讲人简介

Huigang Liang is FedEx Chair of Excellence in MIS at Fogelman College of Business and Economics, University of Memphis. His research interests include socio-behavioral, managerial and strategic information technology issues at both individual and organizational levels in a variety of contexts. His work has appeared in 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 and Journal of Strategic Information Systems, among others. His research has been funded by grants from NIH, FedEx and Sloan Foundation. He has served as Associate Editor for MISQ and ISR and Senior Editor for JAIS. He is currently Senior Editor for MISQ.