Implementación y aplicación de la inteligencia artificial en los procesos de fiscalización tributaria en Latinoamérica.

Abstract

This literature review analyzed the implementation and application of artificial intelligence in tax audit processes in Latin America, particularly in risk detection, tax evasion prevention, taxpayer assistance, and the modernization of fiscal control. The study was non-experimental, descriptive-analytical, qualitative, and documentary in nature; it did not aim to measure or statistically generalize, but rather to describe the state of knowledge on artificial intelligence in tax auditing and critically analyze the relationships, convergences, and gaps among the contributions of the different sources. It was carried out using 25 documents drawn from databases that made up the final matrix, applying the criterion of thematic saturation, understood as the point at which the incorporation of new documents ceases to contribute new categories, nuances, or findings. The results show that AI can support taxpayer segmentation, large-scale data analysis, inconsistency detection, audit prioritization, and preventive auditing practices. The barriers to its adoption and implementation include data quality, structural biases, algorithmic explainability, protection of the exercise of rights, interoperability, technological development, and institutional feasibility. Accordingly, based on the foregoing, artificial intelligence represents an opportunity to modernize the audit function of Latin American tax administrations, under the principles of human oversight, respect for the legal framework, active transparency, public ethics, and respect for due administrative tax process.