Auditors’ Professional Judgment on Sustainability in the Age of AI: From Regulatory and Technical Challenges to Algorithmic Bias
Author:
Maria NICULESCU, Alain BURLAUD
JEL:
M41, M42, Q52
DOI:
10.20869/AUDITF/2025/180/022
Keywords:
generative AI; audit; sustainability; judgement; algorithmic bias; conceptual framework;
Abstract:
The accounting profession, and more specifically that of auditor, is one of the most impacted by the rise of artificial intelligence (AI) and sustainability information regulation. AI is profoundly changing traditional auditing methods, altering the role, approach and responsibilities of auditors, while requiring new skills.
This transformation is particularly striking in the specific field of sustainability auditing, which is becoming increasingly important in a context of heightened demands for corporate transparency and ESG accountability. AI enables the automation and rapid processing of large volumes of data from reporting or external databases, freeing auditors from repetitive tasks. In theory, this automation should allow them to refocus on interpreting results, exercising professional judgement, making critical decisions and managing ESG issues.
However, this new situation raises several major questions, centred on one key issue: what conceptual framework should guide the training of professional judgement by sustainability auditors in a context of regulatory and technological change?
As we do not yet have the necessary perspective on such practices, nor any consolidated empirical data, this article is intended as a conceptual essay aimed at exploring and enriching the existing framework for professional judgement. It proposes a conceptual framework for structuring judgement training in sustainability auditing practices, integrating both enhanced European standards and the disruptive transformations brought about by AI.
The analysis is based on a critical review of academic literature, European and international regulations, and the authors` experience. It is complemented by a qualitative approach using focus groups, which aimed to validate the proposed analysis framework and identify the emerging skills that are essential in this new professional paradigm.
As part of this research, the authors used generative AI (GPT-4 version, 2025) to facilitate documentary research, particularly in the collection of empirical examples illustrating the contributions, technical challenges and algorithmic biases associated with these technologies.
Abstract(273KB)
Article(495KB)