Key insights included:
➡️ Governance & Accountability: Boards and senior management remain accountable for AI systems, outcomes, oversight, risk management and compliance, with AI governance expected to align with the CBUAE Model Management Standards #MMS;
➡️ Fairness & Bias Controls: Institutions must ensure AI systems do not produce discriminatory or manipulative outcomes and should conduct regular testing to identify and remediate bias, including following material model changes;
➡️ Transparency & Explainability: Firms should clearly disclose the use of AI, particularly for high-impact decisions, provide understandable explanations of AI-driven outcomes and communicate with consumers in both Arabic and English;
➡️ Human Oversight: Meaningful human oversight should be applied to AI-driven processes, particularly where decisions materially affect consumers, with customers able to request human review of AI-generated outcomes;
➡️ Continuous Monitoring: AI models should be subject to ongoing monitoring, validation and periodic independent review to ensure reliability, fairness and continued alignment with consumer protection objectives; and
➡️ Third-Party AI Risk Management: Firms using external AI providers should conduct due diligence, maintain audit rights, monitor provider developments and ensure outsourced AI solutions meet the same standards as internal models.
✅ Firms should review AI governance frameworks, model risk management controls, consumer protection processes and third-party AI oversight arrangements to ensure alignment with the CBUAE’s expectations for responsible AI adoption and use.
💡 Plenitude’s AI Assurance Services are backed by world-leading experts, supporting clients in the ongoing monitoring and assessment of risks and controls. Visit our website or reach out directly to learn more: https://www.plenitudeconsulting.com/services/ai-assurance-services