A multi-layered regulatory approach

National AI Strategy 2031

Launched in 2017, the UAE National AI Strategy positions AI as a national priority across 8 sectors. It establishes the UAE's ambition to become a global AI hub with government-led adoption and innovation.

Federal Data Protection (PDPL)

Federal Decree-Law No. 45 of 2021 governs data processing including AI systems. Covers consent, purpose limitation, and data subject rights. Penalties up to AED 5,000,000 for severe breaches.

Dubai AI Seal

The Dubai Centre for AI's (DCAI) AI Seal Verification System (launched January 2025) provides a six-tier trust-mark certification for AI applications meeting ethical, safety, and transparency standards across public and private sectors.

CBUAE AI Guidance

The February 2026 Guidance Note on Consumer Protection and Responsible AI applies to Licensed Financial Institutions (LFIs) supervised by the Central Bank. It sets expectations for the right to human review of AI-driven decisions, board accountability for AI outcomes, and explainability of AI logic.

AI Ethics Principles

Digital Dubai's 4-pillar ethics framework — Ethics, Values, Principles, and Guidelines — provides non-binding guidance on AI transparency, fairness, accountability, and human oversight.

Evolving Enforcement

Multiple authorities oversee AI governance — UAE Data Office, Digital Dubai, CBUAE, and DFSA. DIFC Regulation 10 and the Federal PDPL are binding; the National AI Strategy and Dubai AI Ethics Principles are non-binding guidance. Sector-specific requirements continue to evolve.

AI governance under Regulation 10

AI Impact Assessment (AIIA)

Required under Article 29 (read with Regulation 10) before deploying AI systems that process personal data. Must cover data flows, model architecture, bias risk, and proportionality of automated decision-making.

Transparency & Explainability

Data subjects must be informed of AI processing, the logic involved, and the significance of automated decisions. Right to meaningful explanation of outcomes.

Human Override

Under Article 11 of DIFC Law No. 5, data subjects have the right to contest automated decisions. Regulation 10 requires meaningful human oversight integrated into system design.

Algorithm Fairness Testing

Regular bias audits for AI systems making decisions about individuals. Commissioner notification required for high-risk processing activities.

Financial services AI governance

Board Accountability

The board of directors is directly accountable for AI outcomes. Requires documented AI strategy, risk appetite statement, and regular board-level AI risk reporting.

Model Risk Management

AI models must be integrated into the firm's Model Risk Management framework. Full lifecycle governance from development through validation, deployment, and retirement.

Third-Party AI Outsourcing

Enhanced due diligence for third-party AI providers. Contractual requirements for transparency, auditability, and business continuity. Material outsourcing notification to DFSA.

Stress Testing

AI models used in critical business functions must undergo stress testing and scenario analysis. Results integrated into firm-wide risk reporting.

Central Bank AI requirements for financial institutions

Credit Decision Reason Codes

CBUAE guidance expects AI-driven credit decisions to include clear explanations. Consumers should be able to understand the basis for denial or limitation of financial services.

Regular Bias Monitoring

CBUAE responsible AI guidance encourages financial institutions to monitor AI models for fairness and bias, with documented remediation plans for identified disparities.

Consumer Protection

CBUAE guidance expects AI systems interacting with consumers to provide clear escalation paths to human agents, with appropriate response times for AI-related complaints.

Model Inventory & Documentation

Complete inventory of all AI models in production, including purpose, data sources, validation status, and responsible business owner. Available for supervisory inspection.

DIFC Data Protection vs GDPR

RequirementDIFC Data Protection LawEU GDPR
Breach notificationAs soon as practicable to Commissioner (Art. 42)72 hours to supervisory authority
Private right of actionYes — via DIFC CourtsYes — via national courts
Data export adequacyCommissioner's adequacy list (includes EU/EEA)European Commission adequacy decisions
DPO requirementRequired for high-risk systematic processing (Art. 16)Required for public bodies & large-scale monitoring
AI-specific provisionsRegulation 10 — autonomous & semi-autonomous systemsLimited (Art. 22 automated decisions)
PenaltiesUp to USD 100,000 per violation (Schedule 2)Up to €20M or 4% global turnover

How Utisha maps to UAE regulations

Regulatory RequirementUtisha Capability
CBUAE Guidance — ExplainabilityFull AI transaction logging with timestamped provenance, exportable audit packages
DIFC Reg 10 — AI system registerAI Risk Registry with risk-level tagging, automated classification suggestions
UAE AI Guidelines — GovernanceBoard-ready compliance dashboards, periodic report generation
Dubai AI Ethics — TransparencyAutomatic AI content labelling, decision explanation generator
DIFC Reg 10 — AI Impact AssessmentPre-deployment assessment workflow, data flow mapping, bias risk scoring
DIFC Reg 10 — Human overrideConfigurable intervention points, emergency stop, confidence-based escalation
DIFC Reg 10 — Fairness testingBias audit tooling with protected characteristic analysis, disparity metrics
DFSA — Model risk managementModel inventory, validation tracking, lifecycle governance, drift detection
DFSA — Third-party AI oversightProvider-agnostic LLM Gateway with DLP scanning, output validation, cost tracking
CBUAE — Bias auditsAutomated fairness reports, remediation workflow, audit trail
CBUAE — Credit reason codesAdverse action notice generation with compliant reason code mapping

Deployment & Data Residency

Utisha is a self-hosted platform — you deploy it on your own infrastructure, wherever your data residency requirements dictate. For UAE organisations, this means running Utisha on UAE-based infrastructure so your data never leaves the country. Combined with Utisha's provider-agnostic LLM Gateway, you can route AI processing through locally-hosted models such as TII's Falcon — the UAE's sovereign large language model — ensuring full data sovereignty without compromising on capability.

Utisha's managed cloud hosting is currently available in EU regions (EU, DE, FR, NL). UAE deployments are supported through the self-hosted model on your own or co-located infrastructure.

Assess your UAE AI readiness

Book a Compliance Assessment

Or email us directly at info@utisha.com