Assuring Sustainable Futures: Auditing Sustainability Reports using AI Foundation Models

Abstract

Ensuring the validity and reliability of the economic activities communicated in sustainability reports is essential to comply with the exposure draft of the International Standard on Sustainability Assurance 5000. In recent years, a significant technological transformation has been observed, attributable to the far-reaching functionalities associated with Large Language Models (LLMs). This study combined LLMs with sustainability report auditing to develop a sustainability-related audit prompt framework. Using a case study, we demonstrated the usefulness of the developed framework, analyzing sustainability reports to automatically verify and validate the sustainability of economic activities described within them. The integration of LLMs signifies the advent of a new era in auditing of sustainability reports.

Keywords: Generative Artificial Intelligence, Foundation Models, Large Language Models, Sustainability, Auditing, Assurance

JEL Classification: M40, M41, M42, O30, O32, O33, Q55, Q56