
ISO/IEC 42001: the certifiable AI management standard
In short: ISO/IEC 42001 is the international standard for an AI management system: what it requires, how audits use it and what certification does not prove.
ISO/IEC 42001 is the international standard, published on 18 December 2023, that specifies requirements for an AI management system: the policies, roles, processes and controls an organisation uses to develop, provide or use AI responsibly. Its catalogue entry says it applies to any organisation that provides or uses products or services built on AI systems. Because it sets requirements, organisations can be audited against it.
ISO/IEC 42001 in plain words
ISO 27001 is the standard companies point to for how they run information security. ISO 42001 plays the same role for how they run AI.
How it works
It follows the harmonised structure shared by ISO management-system standards. The published preview lists clauses for context, leadership, planning, support, operation, performance evaluation and improvement. Planning and operation include an AI risk assessment, AI risk treatment and an AI system impact assessment. Annex A holds reference control objectives and controls, and Annex B gives implementation guidance for them.
The introduction names what those processes should cover: trustworthiness concerns such as security, safety, fairness, transparency and data quality across the life cycle, plus the management of suppliers and third parties. Internal audit and management review check the system from inside. A certification audit tests it against the same requirements from outside.
Example
As an illustration, an IT services company building AI solutions for banks pursues 42001 certification so that clients' vendor-risk teams can rely on one audited management system instead of sending a new 200-question AI questionnaire for every deal.
Where you'll meet it
- Enterprise procurement and vendor due diligence
- AI and cloud providers that publish their certificates
- Consulting and audit offerings
- Control mappings that set it beside the EU AI Act and the NIST AI RMF
What it changes for you
For services and product companies selling to regulated industries, certification may shorten vendor-risk reviews. For engineering teams, it formalises practices such as risk and impact assessments, documented information and planned changes. For smaller teams, its structure is useful even without certification.
ISO/IEC 42001 vs ISO/IEC 27001
ISO 27001 covers information security management. ISO 42001 covers AI management, including concerns such as fairness, transparency and systems that keep learning in use, which a security standard is not written for.
Common mistake
My view: treating certification as proof that every AI system is safe. It shows the management system works; each system still needs its own risk assessment, evals and monitoring.
Related terms
AI governance · NIST AI RMF · EU AI Act · AI-BOM · Model card
From the archive: Voluntary AI frameworks are regulatory signals
Certification shows the management system works, and each AI system still needs its own risk assessment, evals and monitoring.


