Competition and Diversity in Generative AI

Abstract

Recent evidence, both in the laboratory and in the wild, suggests that the use of generative artificial intelligence reduces the diversity of content produced. As generative AI penetrates more markets and reaches widespread adoption, should we still expect to see content homogenization? Our work begins with a simple observation: content producers do not act in isolation. Instead, they compete with one another (e.g., for customers or attention). We explore the impact competition has on both content diversity and overall social welfare when producers use generative AI. Through a formal game-theoretic model, we show that competitive markets select for diverse AI models, mitigating monoculture. We further show that a generative AI model that performs well in isolation (i.e., according to a benchmark) may fail to provide value in a competitive market. Our results highlight the importance of evaluating generative AI models across the breadth of their output distributions, particularly when they will be deployed in competitive environments. We validate our theoretical results with a series of experiments involving generative text and image models. Overall, our results suggest that homogenization due to generative AI is unlikely to persist under strong competition. This paper was accepted by Itai Ashlagi, market design, platform, and demand analytics. Supplemental Material: The online appendix and data files are available at https://doi.org/10.1287/mnsc.2025.02761 .