AI-Enabled CTAS and Digital Tax-Fraud Detection: A PLS-SEM Study in Indonesia

Abstract

This study investigates the factors determining digital tax fraud based on the New Fraud Star Theory, with great emphasis on the moderating role of AI-empowered CTAS. Data were collected from 107 corporate taxpayers in Indonesia through a structured survey and analyzed using Partial Least Squares Structural Equation Modeling. The results indicated that System Pressure, Technological Capability, and External Digital Pressure significantly heightened fraud attempts, while Digital Opportunity, AI Rationalization, Cyber Arrogance, Internal IT Governance, and Techno-Culture were not significant. The model explained a substantial variance in the effectiveness of fraud detection with R² = 0.723. Moderation analysis showed that AI-powered CTAS significantly weakened the effects of System Pressure (X1×CTAS), Technological Capability (X4×CTAS), Internal IT Governance (X6×CTAS), and External Digital Pressure (X7×CTAS). These findings identify CTAS's strategic role in improving compliance by enabling real-time data integration, anomaly detection rules, and strengthened access control. Implications are that digital governance reforms should give full attention to the establishment of robust AI-empowered monitoring systems to minimize the risk of tax fraud effectively.