UAB Magazine Issue October 2026

85 Study Union of Arab Banks - October - 2026 However, AI is not a substitute for sound governance. A system may produce false alerts, overlook suspicious activity, or generate an explanation that appears convincing but is incorrect. Poor-quality data can lead to biased or unreliable decisions, while the use of external AI providers can expose confidential customer information to additional risks. GenerativeAI creates further challenges as employees may inadvertently share sensitive information with external tools, rely on inaccurate outputs, or use systems that have not been approved by the bank. More autonomous AI systems, sometimes described as agenticAI, raise additional questions about how actions are authorized, monitored, and stopped when something goes wrong. The U.S. banking regulators have increasingly focused on model risk management, governance, third-party providers, data security, and the use of AI in important banking activities. However, it would be inaccurate to suggest that a single comprehensive U.S. banking regulation already governs every form of AI. Existing supervisory frameworks remain relevant, while approaches to newer AI technologies continue to develop. For banks, the lesson is clear: AI adoption should be accompanied by defined responsibilities, appropriate testing, human oversight, and controls proportionate to the importance of the decisions being made. What do these developments mean for banks? The three areas overlap, but they do not create identical risks. A bank that uses AI to improve customer service faces different issues from one that facilitates stablecoin transactions or holds digital assets in custody. A risk-based approach should reflect these differences. Table 1. Emerging technologies and key compliance priorities Area Main opportunity Principal concern Priority for banks Digital assets New services and more efficient settlement Unclear ownership, illicit finance, and legal uncertainty Understand the asset, counterparties, and applicable rules Stablecoins Faster payments and additional transfer channels Reserve quality, redemption, sanctions, and AML/CFT risks Assess issuers, redemption arrangements, and transaction controls Artificial intelligence Better monitoring, fraud detection, and operational efficiency Unreliable outputs, bias, data exposure, and weak oversight Establish governance, testing, and human review Combined applications More integrated and automated financial services Risks spreading across connected systems and providers Assess the complete transaction and technology chain Source: Union of Arab Banks – Research Department based on U.S. regulatory guidance and banking- sector analysis.