DATA MANIPULATION RISKS IN ARTIFICIAL INTELLIGENCE-BASED ESG REPORTING

Abstract

In recent years, the increasing importance of sustainability-oriented investment strategies has significantly expanded corporate practices related to environmental, social, and governance disclosures. In this context, the Environmental, Social and Governance (ESG) reporting framework has become one of the most widely used tools for evaluating corporate sustainability performance. With the acceleration of digital transformation, the integration of Artificial Intelligence technologies into ESG reporting processes has grown rapidly, enabling organizations to collect, analyze, and report sustainability-related data more efficiently. Despite these advantages, the adoption of AI-based data analytics systems also introduces new risks concerning data reliability, transparency, and accountability in corporate reporting. One of the major concerns discussed in the literature is the possibility that companies may selectively disclose sustainability information to present their environmental performance more favorably. This phenomenon is commonly described in the literature as Greenwashing. The use of AI-driven reporting systems may further complicate these risks, particularly through algorithmic data selection, automated report generation, and opaque decision-making processes. The dependence of AI models on training data and the limited transparency of algorithmic systems may therefore create challenges in ensuring the accuracy and credibility of ESG disclosures. This study aims to examine potential data manipulation risks associated with AI-based ESG reporting through a comprehensive literature review. By synthesizing the existing literature on ESG reporting, artificial intelligence, and greenwashing, the study discusses how these risks may affect the reliability of corporate sustainability reporting and highlights the importance of transparency, independent verification, and responsible AI governance mechanisms.