ISO IEC 23053-2022 PDF

Full title and description

ISO/IEC 23053:2022 — Framework for Artificial Intelligence (AI) Systems Using Machine Learning (ML). This international standard provides a conceptual framework and shared terminology for describing AI systems that use machine learning, defining components and functions of ML-based AI systems within the broader AI ecosystem to improve clarity, interoperability and governance.

Abstract

ISO/IEC 23053:2022 establishes an AI/ML framework for describing a generic AI system that uses machine learning technology. It defines the system components, functions and relationships needed to describe, document and discuss ML-based AI systems across technical and non‑technical stakeholders, supporting consistent communication and the development of complementary standards and governance practices.

General information

• Status: Published.
• Publication date: 20 June 2022.
• Publisher: ISO/IEC (joint publication by the International Organization for Standardization and the International Electrotechnical Commission).
• ICS / categories: 35.020 (Information technology — IT in general).
• Edition / version: Edition 1 (2022).
• Number of pages: 36 pages (first edition published 2022).

Scope

This document provides a conceptual framework and terminology for AI systems that use machine learning. It is applicable to all types and sizes of organizations (commercial, public sector and not‑for‑profit) that design, develop, deploy or use ML‑based AI systems. The framework aims to make descriptions of ML systems consistent and to support alignment with other standards addressing risk, governance, quality and trustworthiness.

Key topics and requirements

• Definition of concepts and shared terminology for ML‑based AI systems (models, datasets, training, inference, deployment, lifecycle roles).
• Component model describing system elements and their functions within the AI ecosystem (data, models, training pipelines, inference engines, monitoring and maintenance).
• Guidance to support consistent description of ML workflows so other standards (risk, governance, management) can reference a common model.
• Applicability to all ML approaches including classical ML and deep learning; support for describing generative and self‑supervised techniques (subject to later amendments addressing Generative AI explicitly).
• Alignment cues to enable interoperability with AI risk, trustworthiness and management‑system standards rather than prescriptive technical requirements.

Typical use and users

Used by standards writers, system architects, ML engineers, technical and non‑technical product managers, auditors and regulators as a foundational reference to describe ML‑based AI systems consistently. Also useful for procurement, documentation, training material, and to support mapping between technical implementations and governance/risk frameworks.

Related standards

ISO/IEC 23053 is part of the ISO/IEC JTC 1/SC 42 family of AI standards. Closely related and frequently referenced documents include ISO/IEC 22989 (AI — concepts and terminology), ISO/IEC TR 24028 (overview of trustworthiness in AI), ISO/IEC TR 24029‑1 (assessment of neural network robustness — overview), ISO/IEC 23894 (guidance on AI risk management) and management‑system guidance such as ISO/IEC 42001 (AI management systems). These documents together form a complementary set covering terminology, trustworthiness, risk, quality and governance for AI systems.

Keywords

AI framework, machine learning, ML systems, terminology, AI components, model lifecycle, deployment, monitoring, AI governance, SC 42.

FAQ

Q: What is this standard?

A: ISO/IEC 23053:2022 is an international standard that provides a conceptual framework and common terminology for AI systems using machine learning, to support consistent description, communication and alignment with other AI standards.

Q: What does it cover?

A: It covers the high‑level framework and vocabulary for ML‑based AI systems — defining components, functions and relationships across the ML lifecycle (data, models, training, inference, deployment and monitoring) and how those map into the wider AI ecosystem. It is intentionally non‑prescriptive about implementation details, focusing on shared description and alignment.

Q: Who typically uses it?

A: System architects, ML engineers, standards developers, product managers, auditors, procurement and policy teams, and regulators — anyone who needs a shared way to describe ML systems and link them to governance, risk and quality requirements.

Q: Is it current or superseded?

A: ISO/IEC 23053:2022 is a current published international standard (first edition published June 2022). It has active amendment work items (for example a draft amendment addressing Generative AI and other amendments/work items under development), but the 2022 edition remains the published, current edition as of the date shown in this product page. Consult the issuing body for the latest amendment and corrigendum publications before relying on a specific clause for conformity assessment.

Q: Is it part of a series?

A: Yes — it is part of the ISO/IEC JTC 1/SC 42 family of AI standards (foundational terminology, trustworthiness reports, risk guidance, management‑system standards and other technical reports) and is intended to be used alongside those documents for comprehensive AI governance and technical guidance.

Q: What are the key keywords?

A: Keywords include: AI, machine learning, ML framework, model lifecycle, dataset, training, inference, monitoring, terminology, SC 42, trustworthiness, governance.