Generative AI-assisted evaluation of ESG practices and information delays in ESG ratings

被引：0
作者：
Wang, Qishu [1 ]
机构：
[1] Seoul Natl Univ, Business Sch, Finance Dept, 1 Gwanak Ro, Seoul 08826, South Korea
来源：
FINANCE RESEARCH LETTERS | 2025年 / 74卷
关键词：
ESG; Information lag; Artificial intelligence; Herding behavior; PERFORMANCE;
D O I：
10.1016/j.frl.2025.106757
中图分类号：
F8 [财政、金融];
学科分类号：
0202 ;
摘要：
We explore the time-lag effects of Environmental, Social, and Governance (ESG) practices' integration within firms from the NASDAQ 100 index, analyzing how this information is reflected over time through Generative AI. By employing ChatGPT-4 to assess ESG progress from 2011 to 2022 in their annual 10-K reports, we identify significant time-lagged effects between the AI- generated evaluations and the ESG ratings, particularly pronounced in the Environmental and Social pillars among firms with mid-range ESG ratings. Our evidence suggests that such information delay in ESG ratings is plausibly driven by reputational herding among firms in their ESG practices, along with potential greenwashing.
引用
收藏
页数：9
相关论文
共 43 条
12345