Implementation of Artificial Intelligence in Fraud Detection and Prevention Through a Systematic Literature Review and Its Implications for the Financial Sector

The increasing complexity of financial fraud in the Digital Era requires more advanced and adaptive detection methods. This study examines the implementation of Artificial Intelligence (AI) in fraud detection and prevention through a Systematic Literature Review (SLR), addressing a critical issue in financial technology that remains highly relevant to both academic and professional communities. Although AI-based fraud detection has been widely studied, this research provides a distinct contribution by integrating technical effectiveness with regulatory alignment. The SLR systematically analyzes studies from major academic databases such as Scopus, Web of Science, IEEE Xplore, ScienceDirect, and Google Scholar to identify key trends, challenges, and implications for the financial sector. The PRISMA framework is used to screen and evaluate relevant literature, ensuring a comprehensive and structured analysis. VOSviewer is applied to visualize key research trends and topic relationships in AI-based fraud detection. The findings indicate that machine learning and deep learning techniques significantly enhance fraud detection accuracy, surpassing traditional rule-based approaches. Natural Language Processing (NLP) has shown effectiveness in analyzing fraud-related documents, while big data analytics facilitates real-time fraud monitoring. However, challenges persist, including data imbalance, regulatory compliance, and data privacy concerns, which must be addressed for successful AI implementation. This study concludes that an integrated AI framework that combines technological advancements with strong regulatory alignment is crucial for effective fraud detection. Future research should explore empirical case studies and real-world applications to validate these theoretical findings.