Unmasking ESG exaggerations using generative artificial intelligence

Authors

Yunfang Luo, Beijing Institute of Petrochemical Technology, Beijing, ChinaFollow
Tao Yang, Beijing Institute of Petrochemical Technology, Beijing, ChinaFollow
Qingan Li, Hong Kong Baptist University, Hong Kong, ChinaFollow
Qiang Liu, Beijing Institute of Petrochemical Technology, Beijing, ChinaFollow
Xiling Cui, Beijing Institute of Petrochemical Technology, Beijing, ChinaFollow

Abstract

Exaggeration is a major indicator of greenwashing, typified by excessively optimistic or idealistic portrayals for environmental protection. The purpose of this study is to identify exaggerated information in environmental, social and governance (ESG) reports by using generative artificial intelligence (GenAI). We analyze a collection of ESG reports using three prompt engineering strategies: few-shot, zero-shot, and chain of thought (COT). We also cross-validate our results using traditional text analytics and human intelligence. Using this strategy, we evaluate exaggeration in ESG reports in a novel way using GenAI. The use of GenAI creates a strong foundation for further study in these and related fields.

Recommended Citation

Luo, Yunfang; Yang, Tao; Li, Qingan; Liu, Qiang; and Cui, Xiling, "Unmasking ESG exaggerations using generative artificial intelligence" (2024). ICEB 2024 Proceedings (Zhuhai, China). 7.
https://aisel.aisnet.org/iceb2024/7