Artificial Intelligence in Financial Reporting Fraud Detection: An Empirical Investigation of Auditor Trust as a Mediating Mechanism in U.S. Markets

Abstract

Artificial intelligence (AI) is increasingly used to support financial reporting fraud detection, yet algorithmic accuracy alone does not determine how effective AI proves to be in practice. This study develops and tests an integrated model in which Auditor Trust in AI Systems mediates the effects of AI Predictive Capability, Explainable AI, Data Quality, and AI Governance and Internal Control Quality on Perceived Financial Reporting Fraud Detection Effectiveness. The model draws on human–AI trust theory, the automation–augmentation perspective, and agency theory. Data from 450 U.S. auditing and financial reporting professionals were collected by structured questionnaire and analysed using Partial Least Squares Structural Equation Modelling in SmartPLS 4. All four antecedents significantly predicted Auditor Trust in AI Systems (R² = .697), which in turn strongly predicted fraud detection effectiveness (R² = .524; β = .724, p < .001) and significantly mediated the effects of all four antecedents. The findings show that AI-enabled fraud detection is a socio-technical outcome that delivers value only when systems are explainable, built on sound data, embedded in appropriate governance structures and trusted by their users. The study contributes an integrated framework linking technological, informational, organisational and behavioural conditions, with practical guidance for audit firms and regulators. JEL classification numbers: M41; M42; O33; G30. Keywords: artificial intelligence, financial reporting fraud detection, auditor trust, explainable AI, data quality, AI governance.