AI-Enhanced Governance for ESG Reporting Integrity: A Sector-Specific Framework Balancing Algorithmic Detection and Human Judgment

Abstract

Purpose: This paper aims to develop a sector-specific governance framework to understand the application of artificial intelligence
in enhancing the quality of ESG reporting, while recognizing the value of human judgement in high-risk disclosure environments.

Design/Methodology/Approach: This conceptual paper uses a theory-driven qualitative approach. It draws on a critical review
of the literature and discourse analysis of 120 corporate sustainability reports from the energy, financial services and consumer
goods industries.

Findings: Environmental disclosures have the highest verification potential on AI since metrics are more benchmarkable and
standardized. It’s worth is reduced in social and governance reporting where narrative interpretation and judgement in context
are vital. It is also revealed that sectoral conditions have their role as operationally measurable industries would be easier to place
under AI-based review as compared to financially oriented narrative reporting.

Originality/Value: The paper re-conceptualizes AI not as a technical solution but as a limited governance mechanism based on
legitimacy, stakeholder and agency theory. It offers a workable hybrid approach where algorithmic detection is used to complement,
not substitute for, board oversight, assurance and judgement in ESG reporting.