Evaluating the Effectiveness of AI-Powered Fraud Detection in Public Finance Audits

Abstract

This study examines the effectiveness of artificial intelligence (AI)-powered fraud detection systems in enhancing the accuracy and efficiency of public finance audits. The central question investigates how AI tools, including machine learning and natural language processing, can address the limitations of traditional audit methods in detecting complex and emerging fraud schemes. Adopting a systematic literature review methodology, the study synthesizes findings from peer-reviewed research to evaluate technological capabilities, implementation challenges, and governance considerations. Results indicate that AI-based systems improve anomaly detection, reduce manual workload, and enable real-time monitoring, but their impact is moderated by data quality, interpretability, and institutional readiness. The discussion emphasizes the need for transparent algorithms, auditor capacity-building, and policy frameworks to mitigate risks such as bias and over-reliance on automated decision-making. The study concludes that integrating AI into audit processes can significantly strengthen fiscal oversight when aligned with ethical and accountability principles.