Thames Kay Consultancy LLCCBUAE expectations for AI and ML models.
The Central Bank of the UAE expects licensed institutions to govern AI and machine-learning models with the same rigour as any other material risk — and to prove it.
Supervisory guidance, not optional reading.
Through the joint Guidelines for Financial Institutions Adopting Enabling Technologies and its model management standards, the CBUAE has set out how it expects banks and other licensed institutions to govern AI and machine-learning: clear board accountability, a complete model inventory, independent validation, explainable outcomes and protection for customers affected by model-driven decisions. Supervisors increasingly test against these expectations in reviews.
Who is in scope
- Banks and finance companies using AI/ML in credit, pricing or fraud decisions
- Institutions deploying customer-facing AI such as chatbots and robo-advisory
- Firms relying on vendor or group-level models — reliance doesn't outsource accountability
- Functions using AI in AML/CFT monitoring and regulatory reporting
At a glance
What the CBUAE looks for.
Board & senior management accountability
Ownership of AI/ML risk at board level, with documented oversight — not delegated into the data science team.
Model inventory & lifecycle governance
Every material model identified, risk-tiered and governed from development through retirement.
Independent validation
Models reviewed by parties independent of their developers, on a cycle proportionate to model risk.
Explainability & data quality
Institutions must be able to explain model-driven decisions and evidence the quality of the data behind them.
Customer protection & fairness
AI-driven outcomes must be fair, contestable and free of unlawful discrimination — especially in credit and pricing.
Third-party & outsourcing controls
Vendor AI is subject to the same governance expectations, plus outsourcing due-diligence and exit planning.
Model governance that passes supervision.
Readiness assessment
Your model governance measured against CBUAE expectations, with gaps rated and prioritised.
Framework build
Model inventory, tiering, validation cycle and committee structure designed to fit your institution.
Validation & audit support
Independent review of material models and audit-committee reporting that anticipates the supervisor's questions.
This page summarises CBUAE supervisory expectations for general information — it is not legal advice. Verify obligations against current CBUAE rulebooks, standards and guidance.
Central Bank of the UAEKnow where your models stand before the supervisor asks.
Bring your model inventory — or let us help you build one — and we'll benchmark it against CBUAE expectations.
Book a consultation