PREVENÇÃO À LAVAGEM DE DINHEIRO COM FERRAMENTAS DE IA: POTENCIAL, LIMITES E COMPLIANCE REGULATÓRIO

This article analyzes the potential of artificial intelligence (AI) tools in anti-money laundering (AML), exploring their technical applications, regulatory limitations, and ethical-legal implications based on national and international frameworks. The research adopts a bibliographic and documentary approach, grounded in Brazilian legislation (Law No. 9.613/1998, Law No. 12.846/2013, LGPD), U.S. regulations (FCPA, BSA, USA PATRIOT Act), European directives (GDPR, MiCA, EU AI Act), and guidance from OECD, FATF, and ISO/IEC 42001. The study examines the role of explainable AI, the integration of financial compliance with blockchain analytics, and the emerging challenges posed by crypto-assets, decentralized finance (DeFi), and obfuscation techniques such as mixers and chain-hopping. The paper emphasizes corporate criminal liability, particularly the legal and institutional risks of delegating sensitive decisions to AI without human oversight. It concludes with practical guidelines for the secure, auditable, and ethical implementation of AI in corporate integrity programs, underscoring the urgency of a hybrid governance model — combining technological innovation with human accountability.