Overcoming the AI Opacity in ESG Reporting: A Digital Platform-based Knowledge Boundary-Spanning Perspective

Abstract

Environmental, Social, and Governance (ESG) reporting has become increasingly important for organizations after the introduction of EU directives. The development of ESG platform functionality is impeded by the scattered knowledge across different stakeholders and the absence of crisp regulatory standards. Artificial intelligence-based systems, such as algorithms integrated with ESG training, can potentially transform investment by providing precise and relevant information. Adopting an Action Design Research methodology, we use four effective knowledge boundary-spanning (EKBS) mechanisms to illuminate the practices of a team of three actors (a platform owner, a complementor, and a platform user) co-designing an explainable artificial intelligence (XAI) tool for ESG reporting in the context of a multi-boundary digital platform. Our data analysis suggests that using EKBS mechanisms is essential for ensuring explainability and trust in AI-based tools.