Uso de Herramientas de Inteligencia Artificial en la Detección de Evasión Fiscal en el Meta: Percepción de los Microempresarios y Retos Para la Dian

Abstract

Tax evasion is one of the main challenges to the sustainability of the Colombian tax system, affecting the financing of social, health, and infrastructure projects. In the department of Meta, losses from consumption tax evasion are estimated to range between 50 and 60 billion Colombian pesos annually, mainly due to smuggling and commercial informality (Gobernación del Meta, 2024). In response to this issue, the National Tax and Customs Directorate (DIAN) has begun a modernization process based on the use of emerging technologies such as artificial intelligence (AI), big data, and automation, with the purpose of strengthening tax auditing and promoting compliance (DIANColombia, 2025). This research aims to analyze the impact of implementing artificial intelligence tools in detecting tax evasion, based on business perceptions and the challenges faced by DIAN in the department of Meta. Microenterprises represent a fundamental pillar of the Colombian economy, accounting for 95.3% of the business structure with 1,854,086 active companies as of March 2023, according to the Ministry of Commerce, Industry, and Tourism. These productive units generate around 80% of national employment and contribute approximately 40% of the GDP (Flórez & León, 2024). Through the study of the most common evasion mechanisms, the review of DIAN’s strategies, and the analysis of business perceptions, this study seeks to propose strategies to consolidate a more efficient, equitable, and reliable tax system.