THE DIGITAL FRONTIER OF GOVERNANCE: IMPACT OF ARTIFICIAL INTELLIGENCE ADOPTION ON ESG DISCLOSURE QUALITY: EVIDENCE FROM SAUDI VISION 2030

Abstract

This study explores the link between corporate artificial intelligence (AI) adoption and the quality of environmental, social, and governance (ESG) disclosure among firms listed on the Tadawul All Share Index (TASI) from 2018 to 2024. Set within the transformative framework of Saudi Vision 2030, we develop a firm-level AI adoption index using natural language processing (NLP) analysis of annual reports and examine its impact on ESG scores obtained from Refinitiv. Using Agency Theory and Signalling Theory, we suggest that AI acts as a "digital monitor" that reduces information asymmetry, limits opportunistic managerial behaviour, and increases the credibility of non-financial disclosures. To address endogeneity and dynamic panel bias, we utilise the Two-Step System Generalised Method of Moments (GMM) estimator, alongside Propensity Score Matching (PSM) as a robustness check. Findings from a sample of 826 firm-year observations strongly support our hypotheses: AI adoption is positively and significantly related to ESG disclosure quality (β = 5.612, p < 0.01), and this relationship is notably stronger in firms with boards that have higher levels of digital expertise (β_interaction = 4.712, p < 0.01). PSM analysis verifies that firms adopting AI outperform matched non-adopters by roughly 13.2 ESG score points. We also observe that the 2021 introduction of the Saudi Exchange ESG Disclosure Guidelines marks a distinct positive structural break in disclosure practices. These results add to the emerging research on digital governance, extend Agency Theory into the era of algorithms, and have direct policy implications for the Saudi Capital Market Authority (CMA) and comparable regulators in other emerging markets.