Xin Tong, University of Southern California

Abstract

Algorithmic markets, where model providers utilize algorithms to provide products or ser- vices, have become increasingly prevalent in our daily lives. The adoption of more accurate algorithms holds the promise of improving social welfare through enhanced predictive accuracy. However, concerns have been raised about algorithmic monoculture, where all model providers adopt the same algorithms. The prevalence of a single algorithm can hinder social welfare due to the resulting homogeneity of available products or services. In this work, we address the rise of algorithmic monoculture from the perspective of providers’ behavior under competition in the algorithmic market. We propose that competition among model providers could mitigate monoculture, thereby enhancing social welfare in the algorithmic market. By examining the impact of competition on algorithmic diversity, our study contributes to a deeper understanding of the dynamics within algorithmic markets and offers insights into strategies for promoting social welfare in these contexts.
 

Bio

Xin Tong is an associate professor at the University of Southern California.  Previously, he was an instructor in Statistics at the MIT math department.  Dr. Tong studied at Princeton ORFE, advised by Profs Jianqing Fan and Philippe Rigollet. He studies asymmetric classification, social learning and networks, AI ethics, and societal Impacts of AI. He is an associate editor for the Journal of American Statistical Association and the Journal of Business and Economic Statistics.