ETHICAL BOUNDARIES OF ARTIFICIAL INTELLIGENCE USE IN THE ACCOUNTING PROFESSION: AN INTEGRATIVE REVIEW
DOI:
https://doi.org/10.24071/icebmr.v5i1.372Keywords:
Artificial intelligence, Accounting profession, Professional ethics, Human-AI collaboration, Ethical boundaries.Abstract
Artificial intelligence (AI) is increasingly transforming accounting practice by extending automation beyond routine processing toward analytical, predictive, and knowledge-intensive activities. While these developments offer significant efficiency and decision-support benefits, they also raise important ethical and professional concerns related to reliability, transparency, bias, confidentiality, professional judgment, and accountability. This study conducts an integrative review of recent literature to examine how AI is currently applied within the accounting profession, identify the ethical and professional issues arising from its use, and determine the boundaries and safeguards required for responsible AI integration. Using a source-based Scopus search, 169 publication sources were identified and screened, resulting in 41 potentially relevant studies and 12 studies included in the final review. A bottom-up thematic synthesis identified five higher-order themes: task-dependent allocation of human and AI roles; reliability, verifiability, and traceability of AI-generated outputs; non-delegable professional judgment and accountability; data governance, confidentiality, fairness, and transparency; and preservation of professional competence and skepticism. The findings indicate that AI is more suitable for structured, repetitive, and independently verifiable tasks, while greater human involvement is required as tasks become more judgment-intensive and consequential. Based on these findings, the study proposes a three-zone ethical boundary consisting of delegable AI-supported tasks, controlled AI augmentation, and non-delegable professional responsibility. The study concludes that responsible AI use in accounting should be understood as accountable augmentation rather than unrestricted substitution, with the degree of AI autonomy decreasing as the need for professional judgment, contextual interpretation, verification, and accountability increases.
Downloads
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Veronika Galuh Fajar Kartika (Author)

This work is licensed under a Creative Commons Attribution 4.0 International License.