Position Paper: Explainable AI (XAI) in Financial Crimes Detection

Online financial transactions have lowered barriers for criminals to propagate and conceal financial crimes such as fraud and money laundering. In response, technology companies, financial institutions and governments have invested heavily in AI-driven systems that detect and investigate malicious transactions. Despite these investments, financial crime detection systems continue to suffer from high false-positive rates and limited forms of explainability that fall short of stakeholders' operational needs, providing insufficient support for the humans who rely on these systems to make decisions. This position paper argues that explainability in financial crime detection should be treated as a stakeholder-sensitive design challenge rather than an add-on feature. We contend that effective XAI must support stakeholders in understanding how a system reaches its decisions in ways that align with their operational workflows. Finally, we propose future research directions for designing XAI that meaningfully support stakeholders engaged in or affected by high-stakes decision-making in financial crime detection.