BIMSA >
BIMSA Digital Economy Lab Seminar
BIMSA Digital Economy Lab Seminar
Generative AI Meets Data Quality: Innovation or Risk?
Generative AI Meets Data Quality: Innovation or Risk?
Speaker
Longtian Zhang
Time
Friday, January 9, 2026 3:00 PM - 4:00 PM
Venue
A3-2-303
Online
Zoom 435 529 7909
(BIMSA)
Abstract
The widespread adoption of Generative AI raises concerns about potential risks, particularly those arising from excessive reliance on AI. This paper examines both the benefits and drawbacks of this emerging technology through the lens of data quality. We develop a semi-endogenous growth model in which production depends on two types of data: AI-generated data and producer data, the latter representing real-world information. Although AI-generated data are substantially cheaper to produce, their use involves a trade-off in the form of lower data quality, which leads to higher error rates in production. Our analysis shows that firms, operating under competitive equilibrium, tend to underutilize both types of data relative to the optimal allocation. We further demonstrate that, while multiple Generative AI firms exist in the market, the optimal number is one. These findings support the case for government intervention in the AI industry.
Speaker Intro
Longtian Zhang is an Associate Professor of Economics at the Central University of Finance and Economics. He holds a Ph.D. from Tsinghua University. His research focuses on macroeconomics, economic growth, and digital economics. Currently, his projects investigate theoretical models on data and AI economy, and he is the principle investigator of National Natural Science Foundation of China (NSFC) Original Exploration Program and Young Scientists Fund (C category). His work has been published in leading journals such as Management Science and Journal of Economic Dynamics and Control.