Explainable Artificial Intelligence for Protecting the U.S. Financial System from Sanctions Evasion and Trade-Based Money Laundering

Abstract

The integrity of the U.S. financial system is increasingly threatened by sophisticated sanctions evasion networks and trade-based money laundering schemes that exploit cross-border payment channels and global supply chains. While machine learning techniques have enhanced the detection capabilities of anti-money laundering and sanctions screening systems, their limited transparency poses significant challenges for regulatory compliance, supervisory review, and effective financial intelligence generation. This study proposes an explainable artificial intelligence framework for protecting the U.S. financial system by enabling detection models that not only identify illicit financial risk but also clearly articulate the underlying drivers of that risk. The research develops explainable AI-based anti-money laundering models tailored to U.S.-touch payment and trade finance environments, including cross-border wire transfers, correspondent banking flows, and ISO 20022-aligned messaging. The framework targets two priority threat domains identified in U.S. national-security guidance: sanctions evasion involving restricted jurisdictions and proxy counterparties, and trade-based money laundering linked to supply-chain manipulation, over-invoicing, and under-invoicing practices. Explainability mechanisms are integrated at the feature, counterparty, transaction routing, and documentation levels to generate investigator-readable narratives and examiner-ready audit trails.
Empirical results demonstrate that the incorporation of explainable artificial intelligence significantly reduces false-positive alert volumes and shortens investigation time compared to traditional opaque models, while improving the clarity and consistency of suspicious activity reporting. By enhancing transparency, accountability, and operational efficiency, the proposed approach strengthens regulatory trust and supports higher-quality financial intelligence outcomes. These findings underscore the role of explainable artificial intelligence as a critical enabler of effective anti-money laundering enforcement, sanctions compliance, and national-security protection within the U.S. financial system.