ISO 42001
ISO 42001: The Complete Guide to AI Management Systems
21 June 2026 · Updated 7 Nov 2026
ISO 42001, published as ISO/IEC 42001:2023, is the first international standard for an AI management system. It sets out how an organisation governs the AI it builds or buys, through clauses 4 to 10 and 38 Annex A controls, with an AI system impact assessment at its core. It is certifiable, risk based, and not yet mandatory in Australia.
What is ISO 42001?
A lot of AI governance discussion stays at the level of slideware, while ISO 42001 is the part an auditor can hold an organisation to. Published by ISO and IEC in December 2023, ISO/IEC 42001:2023 is the world’s first certifiable management system standard for artificial intelligence. It asks an organisation to build, run and improve an AI management system (AIMS), the same way ISO 27001 asks for an information security one: a documented, auditable system for governing AI across its life cycle, reviewed and improved over time, not a one off checklist.
An accredited body can certify against it, which sets it apart from an assessment such as IRAP that produces a report rather than a certificate. What the certificate covers needs to be read carefully, though. Certification says nothing about how good any single model is, because it certifies the governance around the AI rather than the accuracy of the output. For whether it is required here, see is ISO 42001 mandatory in Australia, and for how it compares with information security, ISO 42001 vs ISO 27001.
Who needs ISO 42001?
Any organisation that develops, provides or uses AI is in scope, not only the labs that train models. A bank running a third party model for credit decisions, a SaaS company shipping an AI feature and a government team deploying a chatbot each carry responsibility for how the AI behaves once it reaches real people, so each is covered. The standard applies whether an organisation trains its own models, embeds someone else’s, or switches on AI features in tools it already runs.
The standard is deliberately broad on this, and in Australia the demand is commercial and procurement driven. Customers, partners and tender panels have started asking how an organisation governs AI before they trust it with their data, and a certificate is the cleanest way to answer. Why AI governance matters now sets out what is pushing that change, and ISO 42001 for AI product companies covers the product team view.
What does ISO 42001 require?
ISO 42001 requires two things: a management system, and a set of AI controls to justify. Clauses 4 to 10 carry the management system (context, leadership, planning, support, operation, performance evaluation and improvement), using the Annex SL harmonised structure it shares with ISO 27001 and ISO 9001. Annex A then lists 38 controls across nine objectives. Not all 38 apply by default. They are selected against the organisation’s AI risk and impact assessment, with every inclusion and exclusion recorded, with reasons, in a Statement of Applicability. Annex B holds the implementation guidance for each control, and Annex C lists potential AI related organisational objectives and risk sources.
| Annex A objective | What it governs |
|---|---|
| A.2 Policies related to AI | An AI policy set by leadership, aligned to the organisation’s objectives and risk appetite |
| A.3 Internal organisation | Roles, responsibilities and reporting lines for AI, including who owns each decision |
| A.4 Resources for AI systems | The data, tooling, compute and people an AI system depends on, documented and managed |
| A.5 Assessing impacts of AI systems | The AI system impact assessment: consequences for individuals, groups and society |
| A.6 AI system life cycle | Responsible design, development, testing, deployment and retirement of AI systems |
| A.7 Data for AI systems | How training and operational data is sourced, prepared and governed for quality and provenance |
| A.8 Information for interested parties | What users, regulators and affected people are told about the AI and its limits |
| A.9 Use of AI systems | Responsible operation, including human oversight of automated decisions |
| A.10 Third party and customer relationships | Allocating responsibility across suppliers, model providers and customers |
The selection and the Statement of Applicability are where readiness work concentrates. Our ISO 42001 readiness checklist walks the clauses and the controls in order.
What is an AI system impact assessment under ISO 42001?
This is the requirement that separates ISO 42001 from a security standard. Information security risk asks what could happen to the organisation, while ISO 42001 also asks what an organisation’s AI could do to the people on the other side of it. The standard requires an AI system impact assessment in planning (clause 6) and its execution in operation (clause 8): a documented look at the potential consequences of an AI system for individuals, groups and society, covering fairness and bias, transparency, safety and accountability.
The shift is the point. A security control asks whether the system could be breached, while the impact assessment asks whether the system, working exactly as designed, could harm the people it touches. This is a common point of confusion, because teams run a security risk assessment, file it, and assume the impact assessment is done. It is a separate artefact, and an auditor will ask for it by name. We set out the method in AI risk assessment under ISO 42001.
Is ISO 42001 mandatory in Australia?
No. ISO 42001 is not mandated by Australian law, and there is no pass mark for an individual AI model. What has moved is the policy around it. In September 2024 the Department of Industry, Science and Resources published a Voluntary AI Safety Standard with 10 guardrails, and consulted on mandatory guardrails for AI in high risk settings. Those guardrails (accountability and governance, risk management, data governance, testing, human oversight, transparency, contestability, supply chain transparency, record keeping and stakeholder engagement) line up closely with ISO 42001, so an AI management system is a practical way to meet them ahead of any obligation.
The direction of mandatory regulation is still moving, so we read the voluntary standard as where the floor is today, not where it stays. The pull right now is commercial: customers, tenders and partners asking how an organisation governs AI before they trust it with their data.
How does ISO 42001 relate to the EU AI Act?
Certification is not compliance. The EU AI Act is binding law, in force from 1 August 2024 and phased in through 2025 to 2027, while ISO 42001 is a voluntary standard. Certifying to ISO 42001 does not by itself make an organisation EU AI Act compliant, and the Act does not require the certificate. What the standard gives is the governance scaffold the Act expects: risk management, data governance, human oversight, technical documentation and ongoing monitoring.
Build the management system once and most of the Act’s process expectations are already in place, so certification is a strong foundation rather than a substitute for legal advice. One is the law, while the other is how an organisation shows it governs AI in an orderly way. We compare them in full in ISO 42001 vs the EU AI Act.
How does ISO 42001 fit with ISO 27001?
It reuses the same backbone. Clauses 4 to 10 are shared with ISO 27001, so an organisation that already runs an information security management system finds the context, leadership, internal audit and management review machinery extends to AI with far less duplication, and adds the AI controls and the impact assessment on top. It is not a security standard with AI attached, though. ISO 27001 protects information, while ISO 42001 governs risks an ISMS never set out to address: bias, autonomy, explainability and the effect of an automated decision on a person.
Run together, the two share evidence and audit effort. For a start from an ISO 27001 base, the ISO 27001 guide shows how that management system is built. For the AI specific security threats that accompany this governance, the 2026 guidance standard ISO 27090 is the companion reference, and ISO 27090 vs ISO 42001 explains how the two divide the work.
How long does ISO 42001 take and what does it cost?
It depends on how many AI systems an organisation runs and how mature its governance already is. As an indicative guide from our own engagements, an AI governance gap assessment and AI system inventory takes 3 to 6 weeks, and full implementation to audit readiness 4 to 9 months. Those are indicative figures rather than a quote. Pricing is not published, because it scales with the number and risk of the AI use cases, not a fixed rate.
Certification, where an organisation chooses to pursue it, follows the ISO three year cycle: a Stage 1 documentation review and a Stage 2 implementation audit by an accredited body, then annual surveillance. In practice the larger cost is rarely the audit. It is building the AI inventory and the governance an organisation should have had before the auditor arrives. We break down the drivers in ISO 42001 cost, and Cybernion’s ISO 42001 readiness service runs from the gap assessment and AI system inventory through to audit readiness.
Frequently asked questions
Is ISO 42001 a certification?
Yes. Unlike an IRAP assessment, ISO 42001 is certifiable by an accredited certification body, on a three year cycle of a Stage 1 and Stage 2 audit followed by surveillance audits. The certificate covers the governance system, not the accuracy of any individual AI model.
Who needs ISO 42001?
Any organisation that develops, provides or uses AI, not only the labs that train models. Demand is commercial and procurement driven, as customers, partners and tender panels ask how an organisation governs AI.
How many controls does ISO 42001 have?
Annex A lists 38 controls across nine objectives, A.2 to A.10. The controls are selected against the AI risk and impact assessment, and each inclusion and exclusion is justified in a Statement of Applicability.
Is ISO 42001 mandatory in Australia?
No. Demand is commercial and procurement driven. The Voluntary AI Safety Standard of September 2024 is voluntary, though mandatory guardrails for AI in high risk settings have been under consultation.
How is ISO 42001 different from ISO 27001?
ISO 27001 protects information. ISO 42001 governs AI, and adds an AI system impact assessment that looks at harm to individuals, groups and society, not only risk to the organisation.
Is ISO 27001 needed before ISO 42001?
No, but it helps. The two standards share clauses 4 to 10, so an existing information security management system gives most of the structure. ISO 42001 adds the AI specific controls and the AI system impact assessment.
Does ISO 42001 make an organisation EU AI Act compliant?
No. The EU AI Act is binding law, while ISO 42001 is a voluntary standard. Certification builds the governance the Act expects, but it does not replace legal compliance, and the Act does not require the certificate.
Written by Gaurav Vikash, an ASD endorsed IRAP assessor and senior cyber security leader with 18 years of experience across Australia, the UK and Asia, including CISO and senior security leadership roles. He holds CISSP, CISA, CISM and CRISC and is an ISO 27001 Lead Auditor and ISO 42001 Lead Implementer, and speaks regularly at industry conferences.
Talk to us. We aren’t always chasing a transaction.
Sources:
- ISO/IEC 42001:2023, Information technology, Artificial intelligence, Management system, December 2023
- Voluntary AI Safety Standard, Department of Industry, Science and Resources, September 2024
- EU AI Act, European Commission, 2024
Last updated: 7 November, 2026
