Every AI Policy Ever · open in the interactive register

Published (non-binding, current)Non-binding · Guidelines / framework

AI Ethics Principles

Record SA-01 · SDAIA-P114E

Issued by
Saudi Data and Artificial Intelligence Authority (SDAIA)
Jurisdiction
Saudi Arabia
Level · region
National · Middle East & North Africa
Published / adopted
2025-05
Language
Arabic, English
Official identifier
SDAIA-P114E

Summary

Saudi Arabia's national AI ethics framework, issued by SDAIA for all who design, develop, deploy, use or are affected by AI systems in the Kingdom. It classes AI risks in four levels (little or no risk; limited; high, needing pre- and post-conformity assessments; unacceptable, not allowed) and sets seven principles (fairness, privacy and security, humanity, social and environmental benefits, reliability and safety, transparency and explainability, accountability and responsibility) with controls for each stage of the AI life cycle, including human oversight of irreversible or life-and-death decisions and no use for social scoring or mass surveillance.

Safety topics

risk-based classificationprohibited practiceshuman oversightevaluation & testingtransparency & labellingaccountability & liability

Status notes

The cover gives the year 2025; the document-control table gives May 2025, document type 'Regulatory Documents', document number SDAIA-P114E and version number 1. SDAIA lists the document among its laws and regulations as 'Principles and Controls of AI Ethics' but links the file as a 'draft of Principles of AI Ethics'; the PDF itself carries no draft marking. Adopting entities, defined as those required to comply with the document, designate a head of entity or chief data officer, a compliance officer and a responsible AI officer for AI ethics; SDAIA develops and updates the framework and measures and monitors compliance directly or through sector regulators. Registration with SDAIA is optional and can earn motivational badges. SDAIA's organisational arrangements (Council of Ministers Resolution No. 292 of 27/4/1441 AH, amended by Resolution No. 195 of 15/3/1444 AH) empower it to regulate the data and AI sectors by setting policies, standards and controls and following up compliance with them. SDAIA's generative AI guidelines (SA-02) and deepfakes guidelines (SA-03) build on these principles.

Sources

Open the official source https://sdaia.gov.sa/en/SDAIA/about/Documents/ai-principles.pdf

Local copy

Archived. Last attempt 2026-10-18 04:06:22 UTC

Official source (third-party copy, not published) archive/SA-01/2026-10-11/official.pdf
Retrieved 2026-10-11 15:42:31 UTC
SHA-256 b1b923d44cb39118049830a35254b686ca4720398cdccef3f1a52cea8218911e

Verification

A
The instrument's own official text, or an official register, was stored and confirms the title, date and status. checked 2026-10-18 · official-url-fetched
Recheck of 18 Oct 2026. Stored: the English PDF from sdaia.gov.sa (50 pages; snapshot of 11 Oct 2026, unchanged on 18 Oct 2026), SDAIA's laws and regulations page (18 Oct 2026), and the files that page links: the same PDF (byte-identical, same SHA-256) and SDAIA's organisational arrangements in Arabic. The PDF confirms the title, SDAIA's role, the scope, the four risk levels, the seven principles with life-cycle controls, the roles and the compliance and optional-registration arrangements, and on page 50 the document-control table with May 2025, document number SDAIA-P114E and version 1 (the previous record read only the cover year). The page lists the document under SDAIA's laws and regulations and labels its link 'draft of Principles of AI Ethics'. The Arabic edition of the principles is not stored. The previous notes cited law-firm commentaries describing earlier versions (1.0 of 2022 and 2.0 of September 2023); none is stored and the 2025 edition does not refer to earlier versions.
Second check: Independent check by Claude Code (claude-sonnet-5-5), 2026-10-18: 5 field(s) supported.

History

  1. 2026-10-18notes: The PDF at the official URL carries the year 2025 on its cover and 'Version Numb → The cover gives the year 2025; the document-control table gives May 2025, docume — rewritten: document control, compliance and roles from the PDF; SDAIA's listing of the document and the 'draft' label of its link from the stored page; SDAIA's power to set AI policies, standards and controls from its organisational arrangements; the law-firm account of earlier versions moved to verification.note2026-10-18-recheck-mena-africa-1 · Claude Code (claude-opus-5-5)
  2. 2026-10-18safety_topics: ["accountability & liability", "human oversight", "transparency & labelling", "e → ["risk-based classification", "prohibited practices", "human oversight", "evalua — added risk-based classification (four risk levels with conformity assessment for high risk) and prohibited practices (unacceptable risk not allowed; no social scoring or mass surveillance)2026-10-18-recheck-mena-africa-1 · Claude Code (claude-opus-5-5)
  3. 2026-10-18summary: Saudi Arabia's national AI ethics framework, issued by SDAIA to support the King → Saudi Arabia's national AI ethics framework, issued by SDAIA for all who design, — rewritten to include the four risk levels and the bar on social scoring and mass surveillance, which the framework sets2026-10-18-recheck-mena-africa-1 · Claude Code (claude-opus-5-5)
  4. 2026-10-18official_identifier: – → SDAIA-P114E — null -> SDAIA-P114E, the document number in the document-control table (the same series as SA-02 and SA-03)2026-10-18-recheck-mena-africa-1 · Claude Code (claude-opus-5-5)
  5. 2026-10-18date_published: 2025 → 2025-05 — 2025 -> 2025-05: the document-control table on page 50 gives 'May 2025'2026-10-18-recheck-mena-africa-1 · Claude Code (claude-opus-5-5)

Related records

SDAIA Generative AI Guidelines Saudi Arabia · Published (non-binding, current)

SDAIA Deepfakes Guidelines Saudi Arabia · Published (non-binding, current)

Saudi draft Global AI Hub Law Saudi Arabia · Draft / proposed

Citation

Saudi Data and Artificial Intelligence Authority (SDAIA). AI Ethics Principles [SDAIA-P114E]. 2025-05. https://sdaia.gov.sa/en/SDAIA/about/Documents/ai-principles.pdf. Status: Published (non-binding, current); verified 2026-10-18 (level A). Local copy archive/SA-01/2026-10-11/official.pdf retrieved 2026-10-11 15:42:31 UTC, SHA-256 b1b923d44cb39118049830a35254b686ca4720398cdccef3f1a52cea8218911e. Every AI Policy Ever, version 1.18 (snapshot 2026-10-18), record SA-01.

Improve this record

Know of an amendment, a status change or an error? Report news or an error on GitHub; the form opens with this record filled in.