Information technology — Artificial intelligence — Objectives and approaches for explainability and interpretability of machine learning (ML) models and artificial intelligence (AI) systems

This document describes approaches and methods that can be used to achieve explainability objectives of stakeholders with regard to machine learning (ML) models and artificial intelligence (AI) systems’ behaviours, outputs and results. Stakeholders include but are not limited to, academia, industry, policy makers and end users. It provides guidance concerning the applicability of the described approaches and methods to the identified objectives throughout the AI system’s life cycle, as defined in ISO/IEC 22989.

Technologies de l'information — Intelligence artificielle — Objectifs et approches pour l'explicabilité et l'interprétabilité des modèles d'apprentissage automatique (AA) et des systèmes d'intelligence artificielle (IA)

General Information

Status
Published
Publication Date
04-Sep-2025
Current Stage
6060 - International Standard published
Start Date
05-Sep-2025
Due Date
11-Nov-2024
Completion Date
05-Sep-2025
Ref Project

Buy Standard

Technical specification
ISO/IEC TS 6254:2025 - Information technology — Artificial intelligence — Objectives and approaches for explainability and interpretability of machine learning (ML) models and artificial intelligence (AI) systems Released:5. 09. 2025
English language
69 pages
sale 15% off
Preview
sale 15% off
Preview
Draft
ISO/IEC DTS 6254 - Information technology — Artificial intelligence — Objectives and approaches for explainability and interpretability of machine learning (ML) models and artificial intelligence (AI) systems Released:6. 03. 2025
English language
68 pages
sale 15% off
Preview
sale 15% off
Preview
Draft
REDLINE ISO/IEC DTS 6254 - Information technology — Artificial intelligence — Objectives and approaches for explainability and interpretability of machine learning (ML) models and artificial intelligence (AI) systems Released:6. 03. 2025
English language
68 pages
sale 15% off
Preview
sale 15% off
Preview

Standards Content (Sample)


Technical
Specification
ISO/IEC TS 6254
First edition
Information technology — Artificial
2025-09
intelligence — Objectives and
approaches for explainability
and interpretability of machine
learning (ML) models and artificial
intelligence (AI) systems
Technologies de l'information — Intelligence artificielle —
Objectifs et approches pour l'explicabilité et l'interprétabilité
des modèles d'apprentissage automatique (AA) et des systèmes
d'intelligence artificielle (IA)
Reference number
© ISO/IEC 2025
All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication may
be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting on
the internet or an intranet, without prior written permission. Permission can be requested from either ISO at the address below
or ISO’s member body in the country of the requester.
ISO copyright office
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: +41 22 749 01 11
Email: copyright@iso.org
Website: www.iso.org
Published in Switzerland
© ISO/IEC 2025 – All rights reserved
ii
Contents Page
Foreword .v
Introduction .vi
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviated terms. 5
5 Overview . 6
6 Stakeholders’ objectives . 6
6.1 General .6
6.2 AI user .7
6.3 AI developer .7
6.4 AI product or service provider .7
6.5 AI platform provider .8
6.6 AI system integrator.8
6.7 Data provider .8
6.8 AI evaluator .8
6.9 AI auditor .8
6.10 AI subject .8
6.11 Relevant authorities . .8
6.11.1 Policy makers .8
6.11.2 Regulators .8
6.11.3 Other authorities .9
7 Explainability considerations throughout the AI system life cycle . 9
7.1 General .9
7.2 Inception .10
7.3 Design and development .10
7.3.1 General .10
7.3.2 Development of the explainability component .10
7.3.3 Explainability’s contribution to development .11
7.4 Verification and validation .11
7.4.1 General .11
7.4.2 Evaluation of the explainability component .11
7.4.3 Explainability’s contribution to evaluation . 13
7.5 Deployment .14
7.5.1 General .14
7.5.2 Deployment of the explainability component .14
7.5.3 Explainability’s contribution to deployment .14
7.6 Operation and monitoring . .14
7.7 Continuous validation .14
7.8 Re-evaluation .14
7.9 Retirement . . 15
8 Property taxonomy of explainability methods and approaches .15
8.1 General . 15
8.2 Properties of explanation needs .16
8.2.1 General .16
8.2.2 Expertise profile of the targeted audience .16
8.2.3 Frame activity of interpretation or explanation.17
8.2.4 Scope of information .17
8.2.5 Completeness .17
8.2.6 Depth . .18
8.2.7 Reasoning path .18
8.2.8 Implicit and explicit explanations .19

© ISO/IEC 2025 – All rights reserved
iii
8.3 Forms of explanation .19
8.3.1 General .19
8.3.2 Numeric .19
8.3.3 Visual .19
8.3.4 Textual . 20
8.3.5 Structured . 20
8.3.6 Example-based . 20
8.3.7 Interactive exploration tools . 20
8.4 Technical approaches towards explainability . 20
8.4.1 General . 20
8.4.2 Empirical analysis .21
8.4.3 Post hoc interpretation .21
8.4.4 Inherently interpretable components .21
8.4.5 Architecture- and task-driven explainability . 22
8.5 Technical constraints of the explainability method. 22
8.5.1 General .
...


FINAL DRAFT
Technical
Specification
ISO/IEC DTS 6254
ISO/IEC JTC 1/SC 42
Information technology — Artificial
Secretariat: ANSI
intelligence — Objectives and
Voting begins on:
approaches for explainability
2025-03-20
and interpretability of machine
Voting terminates on:
learning (ML) models and artificial
2025-05-15
intelligence (AI) systems
RECIPIENTS OF THIS DRAFT ARE INVITED TO SUBMIT,
WITH THEIR COMMENTS, NOTIFICATION OF ANY
RELEVANT PATENT RIGHTS OF WHICH THEY ARE AWARE
AND TO PROVIDE SUPPOR TING DOCUMENTATION.
IN ADDITION TO THEIR EVALUATION AS
BEING ACCEPTABLE FOR INDUSTRIAL, TECHNO­
LOGICAL, COMMERCIAL AND USER PURPOSES, DRAFT
INTERNATIONAL STANDARDS MAY ON OCCASION HAVE
TO BE CONSIDERED IN THE LIGHT OF THEIR POTENTIAL
TO BECOME STAN DARDS TO WHICH REFERENCE MAY BE
MADE IN NATIONAL REGULATIONS.
Reference number
ISO/IEC DTS 6254:2025(en) © ISO/IEC 2025

FINAL DRAFT
ISO/IEC DTS 6254:2025(en)
Technical
Specification
ISO/IEC DTS 6254
ISO/IEC JTC 1/SC 42
Information technology — Artificial
Secretariat: ANSI
intelligence — Objectives and
Voting begins on:
approaches for explainability
and interpretability of machine
Voting terminates on:
learning (ML) models and artificial
intelligence (AI) systems
RECIPIENTS OF THIS DRAFT ARE INVITED TO SUBMIT,
WITH THEIR COMMENTS, NOTIFICATION OF ANY
RELEVANT PATENT RIGHTS OF WHICH THEY ARE AWARE
AND TO PROVIDE SUPPOR TING DOCUMENTATION.
© ISO/IEC 2025
IN ADDITION TO THEIR EVALUATION AS
All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication may
BEING ACCEPTABLE FOR INDUSTRIAL, TECHNO­
LOGICAL, COMMERCIAL AND USER PURPOSES, DRAFT
be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying, or posting on
INTERNATIONAL STANDARDS MAY ON OCCASION HAVE
the internet or an intranet, without prior written permission. Permission can be requested from either ISO at the address below
TO BE CONSIDERED IN THE LIGHT OF THEIR POTENTIAL
or ISO’s member body in the country of the requester.
TO BECOME STAN DARDS TO WHICH REFERENCE MAY BE
MADE IN NATIONAL REGULATIONS.
ISO copyright office
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: +41 22 749 01 11
Email: copyright@iso.org
Website: www.iso.org
Published in Switzerland Reference number
ISO/IEC DTS 6254:2025(en) © ISO/IEC 2025

© ISO/IEC 2025 – All rights reserved
ii
ISO/IEC DTS 6254:2025(en)
Contents Page
Foreword .v
Introduction .vi
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviated terms. 5
5 Overview . 6
6 Stakeholders’ objectives . 6
6.1 General .6
6.2 AI user .7
6.3 AI developer .7
6.4 AI product or service provider .7
6.5 AI platform provider .8
6.6 AI system integrator.8
6.7 Data provider .8
6.8 AI evaluator .8
6.9 AI auditor .8
6.10 AI subject .8
6.11 Relevant authorities . .8
6.11.1 Policy makers .8
6.11.2 Regulators .8
6.11.3 Other authorities .9
7 Explainability considerations throughout the AI system life cycle . 9
7.1 General .9
7.2 Inception .10
7.3 Design and development .10
7.3.1 General .10
7.3.2 Development of the explainability component .10
7.3.3 Explainability’s contribution to development .11
7.4 Verification and validation .11
7.4.1 General .11
7.4.2 Evaluation of the explainability component .11
7.4.3 Explainability’s contribution to evaluation . 13
7.5 Deployment .14
7.5.1 General .14
7.5.2 Deployment of the explainability component .14
7.5.3 Explainability’s contribution to deployment .14
7.6 Operation and monitoring . .14
7.7 Continuous validation .14
7.8 Re-evaluation .14
7.9 Retirement . . 15
8 Property taxonomy of explainability methods and approaches .15
8.1 General . 15
8.2 Properties of explanation needs .16
8.2.1 General .16
8.2.2 Expertise profile of the targeted audience .16
8.2.3 Frame activity of interpretation or explanation.17
8.2.4 Scope of information .17
8.2.5 Completeness .17
8.2.6 Depth . .18
8.2.7 Reasoning path .18
8.2.8 Implicit and explicit explanations .19

© ISO/IEC 2025 – All rights reserved
iii
ISO/IEC DTS 6254:2025(en)
8.3 Forms of explanation .19
8.3.1 General .19
8.3.2 Numeric .19
8.3.3 Visual .19
8.3.4 Textual . 20
8.3.5 Structured . 20
8.3.6 Example-based . 20
8.3.7 Interactive exploration tools . 20
8.4 Technical approaches towards explai
...


1 ISO/IEC TSDTS 6254:2024(E)
2 ISO/IEC JTC 1/SC 42/WG 3
3 Secretariat: ANSI
4 Date: 2024-102025-03-03
5 Information technology — Artificial intelligence — Objectives and
6 approaches for explainability and interpretability of machine
7 learning (ML) models and artificial intelligence (AI) systems

8 DTS stage
Warning for WDs and CDs
This document is not an ISO International Standard. It is distributed for review and comment. It is subject to
change without notice and may not be referred to as an International Standard.
Recipients of this draft are invited to submit, with their comments, notification of any relevant patent rights of
which they are aware and to provide supporting documentation.

© ISO/IEC 2024 – All rights reserved

ISO #####-#:####(X)
2 © ISO #### – All rights reserved

ISO/IEC TSDTS 6254:2024(E:(en)
© ISO/IEC 2025
All rights reserved. Unless otherwise specified, or required in the context of its implementation, no part of this publication
may be reproduced or utilized otherwise in any form or by any means, electronic or mechanical, including photocopying,
or posting on the internet or an intranet, without prior written permission. Permission can be requested from either ISO
at the address below or ISO’s member body in the country of the requester.
ISO copyright office
CP 401 • Ch. de Blandonnet 8
CH-1214 Vernier, Geneva
Phone: + 41 22 749 01 11
EmailE-mail: copyright@iso.org
Website: www.iso.orgwww.iso.org
Published in Switzerland
© ISO/IEC 2024 2025 – All rights reserved
iii
ISO/IEC DTS 6254:2024(E:(en)
Contents
Foreword . vii
Introduction . viii
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviated terms . 6
5 Overview . 6
6 Stakeholders’ objectives . 7
7 Explainability considerations throughout the AI system life cycle . 10
8 Property taxonomy of explainability methods and approaches . 17
9 Approaches and methods to explainability . 26
Annex A (informative) Extent of explainability and interaction with related concepts . 54
Annex B (informative) Illustration of methods’ properties . 57
Annex C (informative) Concerns and limitations . 73
Bibliography . 78

Foreword . vi
Introduction . vii
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 Symbols and abbreviated terms . 5
5 Overview . 5
6 Stakeholders’ objectives . 6
6.1 General . 6
6.2 AI user . 7
6.3 AI developer . 7
6.4 AI product or service provider . 7
6.5 AI platform provider . 8
6.6 AI system integrator . 8
6.7 Data provider . 8
6.8 AI evaluator . 8
6.9 AI auditor . 8
6.10 AI subject . 8
6.11 Relevant authorities . 8
6.11.1 Policy makers . 8
6.11.2 Regulators . 8
6.11.3 Other authorities . 8
7 Explainability considerations throughout the AI system life cycle . 9
7.1 General . 9
7.2 Inception . 9
iv © ISO/IEC 2024 2025 – All rights reserved
iv
ISO/IEC TSDTS 6254:2024(E:(en)
7.3 Design and development . 10
7.3.1 General . 10
7.3.2 Development of the explainability component . 10
7.3.3 Explainability’s contribution to development . 11
7.4 Verification and validation . 11
7.4.1 General . 11
7.4.2 Evaluation of the explainability component . 11
7.4.3 Explainability’s contribution to evaluation . 13
7.5 Deployment . 13
7.5.1 General . 13
7.5.2 Deployment of the explainability component . 13
7.5.3 Explainability’s contribution to deployment . 14
7.6 Operation and monitoring . 14
7.7 Continuous validation . 14
7.8 Re-evaluation . 14
7.9 Retirement . 14
8 Property taxonomy of explainability methods and approaches . 14
8.1 General . 14
8.2 Properties of explanation needs . 16
8.2.1 General . 16
8.2.2 Expertise profile of the targeted audience . 16
8.2.3 Frame activity of interpretation or explanation . 17
8.2.4 Scope of information . 17
8.2.5 Completeness . 17
8.2.6 Depth . 18
8.2.7 Reasoning path . 18
8.2.8 Implicit and explicit explanations. 19
8.3 Forms of explanation . 19
8.3.1 General . 19
8.3.2 Numeric . 19
8.3.3 Visual . 20
8.3.4 Textual . 20
8.3.5 Structured . 20
8.3.6 Example-based . 20
8.3.7 Interactive exploration tools . 20
8.4 Technical approaches towards explainability . 21
8.4.1 General . 21
8.4.2 Empirical analysis . 21
8.4.3 Post-hoc interpretation . 21
8.4.4 Inherently interpretable components . 21
8.4.5 Architecture- and task-driven explainability . 22
8.5 Technical constraints of the explainability method . 22
8.5.1 General . 22
8.5.2 Genericity of the method .
...

Questions, Comments and Discussion

Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.