Sign inRequest a demo
Request a demo

Security & compliance · Saudi Arabia

Edrak and the Saudi Data & AI Authority

SDAIA sets the Kingdom's direction on data and AI, including the principles AI systems are expected to follow. Edrak is designed so those principles can be enforced: human approval points, a full record of every decision, and your choice of model for every task.

SDAIA
National authority for data and AI. Sets data and AI policy, ethics principles and governance.
Applies to
Government entities directly; the AI Ethics Principles and Generative AI guidelines are published for all organisations deploying AI in the Kingdom.
Where Edrak stands
Edrak's controls are mapped to SDAIA's AI Ethics Principles and Generative AI guidelines and we share that mapping. SDAIA issues principles and guidance and does not certify vendors.

SDAIA's principles and guidance

SDAIA publishes principles and guidelines. These four are the ones an AI deployment is measured against.

  • Ethics

    AI Ethics Principles

    Fairness, privacy and security, humanity, social and environmental benefit, reliability and safety, transparency and explainability, accountability.

  • GenAI

    Generative AI Guidelines

    Guidance for government and public use of generative AI: acceptable use, data handling and human review.

  • Adoption

    AI Adoption Framework

    How organisations assess readiness, select use cases and govern deployment.

  • Data

    National Data Governance

    Policies inherited from the NDMO, which sits within SDAIA.

How Edrak supports your SDAIA obligations

The obligations are yours. Edrak is built to make meeting them straightforward, and to be clear about what stays on your side.

  • AI ethics and fairness

    Edrak: Human approval points, full records and model choice let you apply your own AI principles to every workflow.

    You: Adopting the principles, assessing impact and monitoring outcomes.

  • Explainability

    Edrak: Each output carries its sources, the model used and the reasoning steps recorded, so decisions can be reviewed after the fact.

    You: Deciding which decisions need human review before they take effect.

  • Human oversight

    Edrak: Approval steps sit where you set them. Nothing completes without the sign-off you require.

    You: Defining the approval points and who holds them.

  • Models and third parties

    Edrak: Routing rules keep restricted data on open-weight models inside your boundary; hosted frontier models handle the rest.

    You: Approving which models may be used, and for what.

  • Procurement and disclosure

    Edrak: Model provenance, deployment location and sub-processors are documented for your procurement and AI registers.

    You: Registering AI use cases and completing impact assessments.

  • Records

    Edrak: Every action, decision and outcome is logged: who asked, which model ran, what it touched, what came back.

    You: Retention periods and forwarding to your monitoring tools.

For your security and compliance teams

We share a requirement-by-requirement mapping to the SDAIA frameworks, a data-flow diagram for your deployment option, and our security overview, under NDA. We also answer questionnaire items directly.

Talk to our team

SDAIA issues requirements and guidance; it does not certify vendors. This page describes how Edrak's controls support your obligations. Edrak does not claim compliance or certification on your behalf. Confirm applicability with your compliance team.