Unlocking Federated Learning for ESG Reporting: Prioritizing Critical Adoption Challenges in an Emerging Economy Context

Abstract

ABSTRACT The growing focus on environmental, social, and governance (ESG) issues has prompted organizations to explore innovative ways to manage their businesses. It has become imperative for organizations to adopt ESG reporting. The applications of federated learning help to enhance the scalability and credibility of ESG reporting in contemporary settings. However, adopting federated learning for ESG reporting is not straightforward; it involves several complexities. This paper identifies and assesses anticipated challenges. Scholarly research databases, including Scopus and Web of Science, were employed to identify relevant research papers. In the proposed study, an initial pool of 15 challenges was drawn from the literature; four additional challenges were introduced by the expert panel, and one was rejected during the Delphi consensus, yielding a final set of 18 validated challenges. Moreover, these 18 challenges were confirmed using the Pythagorean Delphi technique. Then, the Pythagorean fuzzy AHP method was applied to prioritize anticipated challenges. The results suggest that the lack of ESG data standardization and data quality are the top challenges hindering the adoption of federated learning in ESG reporting. Theoretically, the research contributes to the literature by investigating the role of federated learning in effective ESG reporting. From a practical standpoint, the proposed study provides several actionable insights for stakeholders.