Skip to content
AI governance

ISO 42001: the AI management standard, explained

· 5 min read · Vivid Vision

ISO/IEC 42001 is the first internationally certifiable standard for an artificial intelligence management system (AIMS). Published in December 2023, it defines how an organisation must govern, oversee and continuously improve the AI systems it develops or uses: policies, roles, risk management, impact assessment and controls. It is to AI what ISO 27001 is to information security: an auditable management framework recognised worldwide.

This guide explains what the standard requires, how it is certified and why, alongside the AI Act, it is becoming the de facto standard for demonstrating responsible AI governance.

What an AI management system (AIMS) is

An AIMS (AI Management System) is the set of policies, processes and responsibilities through which an organisation directs its AI activities systematically. ISO 42001 follows the high-level structure (Annex SL) common to ISO 27001 and ISO 9001, which makes it easier to integrate with existing management systems. Its logic is the PDCA continuous improvement cycle (plan, do, check, act) and a risk-based approach.

The key difference from other standards is that 42001 introduces concepts specific to AI: the AI system impact assessment on individuals and society, model lifecycle management, data traceability and human oversight.

What the standard requires

ISO 42001 structures its requirements around these blocks:

  1. Context and interested parties. Understanding the role of AI in the organisation and the expectations of customers, regulators and users.
  2. Leadership and AI policy. Management must set an AI policy and assign clear responsibilities.
  3. Planning and risk management. Identifying and treating the risks of AI systems, including an impact assessment.
  4. Support and operation. Resources, competencies, documentation and operational controls throughout the model lifecycle.
  5. Performance evaluation. Internal audits and management review.
  6. Continuous improvement. Correcting non-conformities and raising the level of maturity.

The standard also includes an Annex A with a catalogue of controls (similar to that of ISO 27001) that the organisation selects and justifies according to its risks: data governance, management of AI suppliers, transparency towards users, among others.

How it is certified

ISO 42001 certification is issued by an accredited certification body after a two-stage audit (document review and on-site audit). Once obtained, it is maintained through periodic surveillance audits. It is not a one-off formality: it certifies that the management system works and improves over time. For many companies, the value lies not only in the certificate, but in the operational discipline that getting there imposes.

ISO 42001 vs AI Act: they do not compete, they complement each other

This is the most common confusion. They are not the same thing, and you do not have to choose:

  • The AI Act is a law that is mandatory in the EU, with penalties of up to €35 M or 7 % of turnover. It tells you what you must comply with.
  • ISO 42001 is a voluntary management standard. It provides the how: the operating framework that, properly implemented, helps you demonstrate and sustain that compliance.

An organisation with an AIMS certified to ISO 42001 is much better prepared for the obligations of the AI Act —inventory, risk management, documentation, human oversight— because it already has them operationalised. The standard does not guarantee automatic legal compliance, but it drastically shortens the distance.

Does my company need it?

ISO 42001 is not legally mandatory, but it stops being optional in practice when:

  • Your customers or tenders begin to demand assurances on the responsible use of AI.
  • Your investors ask for an AI risk framework before investing.
  • You operate in a regulated sector (banking, insurance, healthcare) and need to demonstrate control over your systems.
  • You want a competitive differentiator and a solid foundation for AI Act compliance.

Certification is the outcome, not the starting point. The real value lies in building a management system that lives inside the organisation. At Vivid Vision we operate that framework with you, from the inside and in a vendor-agnostic way, and we leave the knowledge installed in your team. — Rafael Tamames, founder of Vivid Vision.

If you want to know where your organisation stands, our AI governance pre-assessment maps your systems, their risks and the real distance from ISO 42001 and the AI Act.

Frequently asked questions

What is ISO 42001?

ISO/IEC 42001 is the first internationally certifiable standard for an artificial intelligence management system (AIMS). Published in December 2023, it sets out how an organisation must govern, oversee and continuously improve its AI systems through policies, risk management, impact assessment and auditable controls.

Is ISO 42001 mandatory?

No, ISO 42001 is a voluntary standard, not a law. However, more and more customers, tenders and investors demand it as proof of responsible AI use, and implementing it makes compliance with the AI Act easier, which is mandatory in the European Union.

How does ISO 42001 differ from the AI Act?

The AI Act is a mandatory law that defines what the EU requires of AI systems, with penalties for non-compliance. ISO 42001 is a voluntary management standard that provides the operating framework —the how— for governing AI. They are complementary: a certified AIMS makes it easier to demonstrate and sustain compliance with the AI Act.

How does a company get certified to ISO 42001?

Certification is issued by an accredited body after a two-stage audit —document review and on-site audit— and is maintained through periodic surveillance audits. It certifies that the AI management system exists, works and improves over time, not that a one-off requirement was met.


Related reading: The AI Act for companies — what the law requires and its deadlines —, how to build an AI governance framework and what AESIA is, the AI authority in Spain.

Sources: the official standard page ISO/IEC 42001:2023 and the Spanish Agency for the Supervision of Artificial Intelligence (AESIA). This guide is for information only and does not constitute legal or certification advice.

Is your AI already in production without a governance framework?

Start with a pre-assessment: we map your AI systems, their risks and your priorities against the AI Act and DORA. We operate governance with you, from the inside.

Request your pre-assessment
Request your pre-assessment

Choose which cookies you accept. You can change your choice or withdraw your consent at any time from Cookie settings, in the footer of every page.

Needed for the site to work and to remember your cookie choice.

They tell us how many people visit the site, which content is most interesting and how you found us if you write to us. Used by Google Analytics, Microsoft Clarity and Ploy analytics.

Cookie policy