Mechanisms for applying artificial intelligence to reduce tax evasion in Ecuador

Abstract

Tax evasion continues to undermine the capacity of the Ecuadorian State to finance its constitutional obligations, due to evasive practices that exceed traditional tax control methods. The general objective of this work was to analyze, through a systematic review, the use of artificial intelligence as a tool for reducing tax evasion in Ecuador. The research adopted a qualitative approach and employed a systematic review methodology to examine relevant scientific literature published in the last five years. Fifteen academic articles were reviewed that addressed models, techniques, and applications of artificial intelligence in tax control processes. The results showed that models such as machine learning, neural networks, random forest, clustering, and expert systems have improved the detection of evasion patterns, optimized audits, identified high-risk taxpayers, and strengthened tax traceability in various countries. It was concluded that, in the Ecuadorian context, legal mechanisms that allow its integration were identified, although challenges persist regarding algorithmic transparency, data protection, technical oversight, and the absence of specific regulation.