SDAIA's AI Ethics Principles are short to read and easy to agree with. The hard part is turning them into decisions on a real project. This is how we do that when we build for Saudi organisations.
Short answer: the principles ask you to know what your system does, who is accountable for it, what data it uses, and how a person can step in. Each of those becomes a design choice, a document or a test.
The seven principles, as build decisions
SDAIA groups its principles under fairness, privacy and security, humanity, social and environmental benefit, reliability and safety, transparency and explainability, and accountability and responsibility. In practice:
- Fairness. Test answers across the groups your users belong to, including Arabic and English speakers and different dialects. Record what you tested.
- Privacy and security. Map every place personal data enters the system: prompts, documents, logs. Apply the PDPL to all of them, and keep data in the client's environment where possible.
- Humanity. Decide which steps a person must approve. An agent can draft; a person signs.
- Social and environmental benefit. Write down why the system exists and who it helps. If you cannot, do not build it.
- Reliability and safety. Ship an evaluation set with the system and run it before every release. See AI evaluation.
- Transparency and explainability. Show users where an answer came from. Grounded, cited answers are the simplest form of explanation.
- Accountability. Name an owner for the system and keep an audit log of what it did.
A one-page checklist for your next project
- Purpose: one paragraph on what the system does and does not do.
- Owner: the person accountable, by role.
- Data map: sources, personal data, where it is stored.
- Human checkpoints: the actions that need approval.
- Evaluation: the test set, the pass bar, the date of the last run.
- Audit: what is logged and who can read it.
What this is not
This is our reading as builders, not legal advice. For your obligations, check the current SDAIA texts and your own counsel. Our security and compliance page states what we align to and what we are not certified for.
To see how these principles shape an agent in production, read our note on the UNESCO ethics forum in Riyadh.