Docs

Appearance

Use Modulos ​

The sections below tell you how to use the UAE AI Ethics Principles and Guidelines in Modulos.

Before you start ​

The UAE AI Ethics Principles and Guidelines (DEC 2022) of the UAE AI Office are "non-binding" (Section 1.1). For more about the guidelines and their sources, see Non-binding guidelines.

Before you add the UAE AI Ethics Frameworks, do these two checks against the guidelines:

  1. Scope. Check if the AI system is in the scope of the guidelines (Section 1.2).
  2. Type of organization. Find if your organization is an AI developer organization, AI operator organization, or the two (Sections 2.1 and 2.2).

In the guidelines, an AI developer organization and an AI operator organization are types of organization. These types are not the Organization level of Modulos.

Modulos has two Frameworks for the UAE AI Ethics Principles and Guidelines. Each Framework has the name "UAE AI Ethics":

  • MFF-4 at the AI System level. Add it to the AI System record of each AI system in scope.
  • OFF-4 at the Organization level. You add it to the organization AI System.

For the UAE AI Ethics Frameworks, Modulos recommends this structure. The guidelines do not set a structure.

Organization AI Systems

Frameworks at the Organization level, for example OFF-4, attach to the organization AI System of your organization. You cannot make a new organization AI System. If your organization does not have one, contact Modulos before you add OFF-4.

Each Requirement of MFF-4 and OFF-4 has the tag "AI Ethics Role" with the two values "AI Developer Organization" and "AI Operator Organization". Modulos added this tag for scope selection by type of organization. The two values are on each Requirement. Thus, the tag removes no Requirement for an AI developer organization or an AI operator organization.

Requirements ​

For the UAE AI Ethics Frameworks MFF-4 and OFF-4, use the Requirements below.

AI System RequirementOrganization RequirementRequirement namePrinciple of the guidelinesInformation
MRF-16ORF-16ExplainabilitySection 3.4, "Explainable AI"Principle 4
MRF-17ORF-17TransparencySection 3.3, "Transparent AI"Principle 3
MRF-18ORF-18AccountabilitySection 3.2, "Accountable AI"Principle 2
MRF-19ORF-19FairnessSection 3.1, "Fairness"Principle 1
MRF-20ORF-20Privacy preservationSection 3.8, "Privacy Preserving AI"Principle 8
MRF-21ORF-21Sustainability and environmental friendlinessSection 3.7, "Sustainable and Environmentally Friendly AI"Principle 7
MRF-22ORF-22Human centralitySection 3.6, "Human Centered AI"Principle 6
MRF-23ORF-23Robustness, safety and securitySection 3.5, "Robust, Safe and Secure AI"Principle 5

For the UAE AI Ethics Frameworks, Modulos recommends the sequence of work below. The guidelines do not set a sequence.

  1. Do the two checks. Do the two checks in Before you start.
    • If the AI system is not in the scope of the guidelines, the guidelines do not apply to it (Section 1.2).
    • Your organization can use the Frameworks for this AI system. That use is a decision of your organization.
  2. Add the Frameworks. Add OFF-4 to the organization AI System. Add MFF-4 to the AI System record of each AI system in scope. See Register an AI System.
  3. Do the work of each Requirement. In Requirements, find the principle of each Requirement. Then read the guidelines of that principle.
    • Find the guidelines for your type of organization. A guideline can name AI developers, AI operators, or the two.
    • Some guidelines do not name an organization. Read these guidelines too.
    • Keep the condition of each guideline, for example "where possible", a critical decision, or "life and death decisions".
    • Do not change "should consider" or "could consider" into "should".
  4. Do the Review of each Requirement.
    • The Requirement Owner examines the Controls and can mark the Requirement as Fulfilled.
    • Controls do not make a Requirement Fulfilled without this Review.
    • See Reviews & Statuses.
  5. Do the work again when the AI system, its use, or its results change. The guidelines set no interval for this work. See the guidelines on continued work in Principle 2 and Principle 5 (Sections 3.2.2.2(e), 3.5.2.1 and 3.5.2.3(a)).

Complete the work ​

Use the UAE AI Ethics Frameworks in Modulos to record your work on each Requirement. A Fulfilled Requirement does not show that all work recommended by the guidelines is complete. Compare your work with the guidelines of each principle.

Check the applicable guidelines for external audits, information for AI subjects, and human control of decisions:

  • External audit and release of results — See Principle 2 for critical decisions and the condition "of civic interest" (Section 3.2.2.3(a)-(b)).
  • Information for AI subjects — See Principle 3 for information about significant decisions (Section 3.3.2.2(a)). For procedures to challenge decisions, see Principle 2 (Section 3.2.2.4).
  • Checks by a person — For critical decisions that are "life and death decisions", see Principle 2 (Section 3.2.2.3(c)).
  • Human case evaluators — For the use of human case evaluators to examine challenges, see Principle 2 (Section 3.2.2.4(d)).
  • Human control — For decisions "related to lethal force, life and death", see Principle 6 (Section 3.6.2.3(a)).

Sources ​

  • AI Ethics: Principles & Guidelines — UAE AI Office, DEC 2022. Section 1.5 names "the UAE AI and Blockchain Council as its owner". Sections 1.1, 1.2, 1.5, 2.1, 2.2, 2.6, 2.11, and 3.1 to 3.8. PDF.

For more information about the UAE AI Ethics Principles and Guidelines and about Modulos, see:

UAE AI Ethics principles

The decision tests, the types of organization, and the eight principles

UAE AI Ethics questions

Short information about the guidelines and the Frameworks in Modulos

Governance operating model

Requirements, Controls, Evidence, and Reviews in Modulos

Evidence

How to attach Evidence to a Control

Control overlap

The Controls that two or more Frameworks use

Disclaimer

This page gives general information. It is not legal advice.