Exploring the moderating role of natural language between the use of AI and auditing and fraud detection in accounting information system: an empirical study in Indonesia

Abstrak

This study aims to investigate the moderating role of Natural Language Processing (NLP) in the relationship between AI-empowered accounting information systems and audit and fraud detection. The research method used is quantitative analysis with data collection through questionnaires distributed to respondents from finance and accounting departments of companies in Indonesia. This study uses multiple regression analysis and Moderated Regression Analysis (MRA) to test hypotheses. The results show that AI in accounting information systems has a significant effect on audit and fraud detection, with prevention and investigation dimensions as the main contributors. NLP partially moderates the relationship between AI and audit and fraud detection, where NLP significantly strengthens the prevention dimension, negatively moderates the investigation dimension, but does not moderate the dimensions of data gathering, data analysis, risk assessment, and detection. Theoretical Contribution: This study extends the literature on AI and NLP integration in accounting information systems by showing that NLP effectiveness is context-specific and differential depending on the AI dimension being moderated. Practical Contribution: These findings provide guidance for audit practitioners and organizations in prioritizing NLP implementation in preventive audit systems, as well as providing careful considerations in implementing NLP for fraud investigation.