Artificial intelligence in risk-based auditing of large enterprise taxation

Abstract

The article examines the theoretical and methodological foundations for applying artificial intelligence (AI) in risk-based audits of large enterprise taxation. The study’s relevance stems from the significant fiscal role of large taxpayers, the complexity of their business transactions, the prevalence of cross-border transactions and transfer pricing practices, and the need to identify tax risks promptly amid the digital transformation of tax administration. The study aims to identify areas for AI application to improve the effectiveness of identifying, assessing, and prioritizing tax risks for large enterprises. Based on an analysis of academic research and international approaches, the study systematizes the key tax risks inherent in large enterprises, including those related to transfer pricing, profit shifting across jurisdictions, transactions involving intangible assets, intra-group financing, corporate income taxation, value-added tax (VAT), and tax compliance. The study explores the potential of machine learning, anomaly detection algorithms, predictive analytics, natural language processing (NLP), cluster analysis, and network analysis for processing large volumes of financial and tax data. This study develops a structural and logical model for applying AI in risk-based audits of large enterprise taxation, integrating information and analytical support, AI tools, tax risk profiling, and the auditor’s professional judgment. AI-driven analysis should provide an analytical basis for identifying areas of heightened tax risk and prioritizing audit procedures, while final audit decisions should remain grounded in the auditor’s professional judgment. The practical significance of the findings lies in the potential application of the proposed approach to enhance the validity and effectiveness of risk-based audits of large enterprise taxation. Keywords: artificial intelligence, risk-based audit, tax audit, large enterprises, tax risks, transfer pricing, machine learning, tax compliance.