ISO/IEC 23053 is the framework and terminology standard for describing AI systems that use machine learning. It names the components, approaches and pipeline stages; it sets no certifiable requirements.
ISO/IEC 23053:2022 is an International Standard from ISO/IEC JTC 1/SC 42. It gives an organisation a shared way to describe an ML-based AI system: the task the model is asked to solve, the model itself, the data it learns from and the tools around it, and the pipeline from data acquisition to operation. It is descriptive. It contains no management system requirements and no conformity assessment route, and no one can be certified to it. Its practical role is to be the common vocabulary behind other standards. ISO/IEC 42001 cites it in its bibliography, and AI governance teams use it to write the AI system description that a 42001 audit samples.
Is it a product standard?
No. It describes the AI system that uses machine learning, not a product or a management system. The obligations live in other standards and in the law.
Who uses it?
Governance leads, architects, data scientists and auditors who need one vocabulary for the system. It is written to be readable by non-specialists.
Where this comes from
The clause record states each unit in our own words and cites the source. The amendment page separates the public draft summary from the published base edition. The base record is ISO standard 74438.
Not legal advice; does not certify. The amendment wording is not held and is not stated.