Title:	CAN GENERATIVE ARTIFICIAL INTELLIGENCE (AI) ASSISTANTS' EVALUATION OF ENVIRONMENTAL, SOCIAL, AND GOVERNANCE (ESG) PERFORMANCE REPLACE PROFESSIONAL EVALUATION?
Author(s):	Victor K. Y. Chan
ISBN:	978-989-8704-62
Editors:	Paula Miranda and Pedro Isaías
Year:	2024
Edition:	Single
Keywords:	Artificial Intelligence, AI, Environmental, Social, Governance, ESG
Type:	Full
First Page:	301
Last Page:	308
Language:	English
Cover:
Full Contents:	click to dowload
Paper Abstract:	This article explores how evaluation of companies'/stocks' environmental, social, and governance (ESG) performance by generative artificial intelligence (AI) compares with traditional, proprietary, professional evaluation, and whether the former is able to replace the latter. The generative AI assistant utilized in the underlying study was Microsoft Copilot, which was requested to accord rating scores to the three individual ESG components, namely, (1) Environmental, (2) Social, and (3) Governance of the top 40 companies/stocks among the S&P 500. The traditional, proprietary, professional evaluation of the companies'/stocks' ESG performance for these three components adopted in this article was the rating scores by Sustainalytics of Morningstar. The correlation coefficient between Copilots' rating score for each of these three components over the top 40 companies/stocks and the corresponding rating score from Sustainalytics was computed. Subsequently, multiple regression of Sustainalytics's ESG Risk Rating score (i.e., a summary ESG score from Sustainalytics as the dependent variable) on Copilot's rating scores for the three components above (as the independent variables) over the top 40 companies/stocks was performed. It was found that the correlation coefficients were respectively -.576 (p = 0.000), -.166 (p = .306), and -.171 (p = .291). The multiple regression included Copilot's rating scores for all the three components above as the independent variables with the R2 = .441, the F-test's F statistic = 9.486 (df = (3, 36) and p = 0.000), the respective regression coefficients being -5.886, 2.176, and -2.185 and the corresponding t-tests' t values being -3.289 (p = 0.002), .955 (p = .346, and -1.075 (p =.290).
	Go Back