Modelos predictivos de inteligencia artificial para prevención de lavado: Análisis comparativo tripartito en la región andina

Money laundering constitutes a critical threat to financial stability in Ecuador, Colombia, and Peru, countries facing unique challenges arising from drug trafficking and informal economies representing 37% to 77% of GDP. This research analyzes the potential of artificial intelligence predictive models to strengthen regional anti-money laundering systems through systematic comparative study. Findings reveal that, although technologies like Graph Neural Networks and XGBoost achieve accuracies exceeding 99%, none of the 68 analyzed studies specifically addresses the Andean context. The review identifies critical gaps in adaptation to informal economies, cross-border cooperation, and limited computational resources. Results suggest that ensemble models like Random Forest and XGBoost offer favorable balances between effectiveness and regional viability, while cross-border cooperation could generate superior detection benefits compared to independent national efforts. Successful implementation requires incremental strategies considering local socioeconomic particularities, secure data exchange frameworks, and simultaneous institutional strengthening to maximize the effectiveness of advanced technological systems.