Impact of an Interface Metaphor Based on Algorithmic Thinking on Academic Performance in Public Accounting Students

Abstract

Financial fraud has become one of the main challenges in the financial and accounting sectors due to the increasing sophistication of fraudulent schemes and the rapid digitalization of financial services. In this context, artificial intelligence (AI) has emerged as a strategic solution for improving fraud detection through predictive analytics, anomaly detection, and automated decision-making. This study presents a systematic literature review (SLR) aimed at analyzing the main artificial intelligence techniques applied to financial fraud detection, as well as their benefits, challenges, and future research directions, with particular emphasis on the Colombian context. The review followed the methodological guidelines proposed by Barbara Kitchenham and included studies published between 2020 and 2025 in databases such as Scopus, ScienceDirect, Google Scholar, and Redalyc. A total of 388 records were initially identified, from which 40 studies were selected after applying screening and quality assessment criteria. The findings indicate that machine learning, deep learning, graph-based models, and hybrid architecture are the most widely used approaches for detecting fraud in banking systems, Fintech ecosystems, tax administration, and public procurement processes. The results also reveal significant improvements in anomaly detection, operational efficiency, and real-time monitoring. However, important challenges remain regarding explainability, data accessibility, regulatory frameworks, and the shortage of specialized human talent. Overall, the study highlights the growing relevance of AI-driven fraud detection systems and the need for explainable, scalable, and ethically aligned solutions for future financial environments.