ISO/IEC 42001 guide
ISO 27001 to ISO 42001: how to extend your ISMS into an AI management system
Clause by clause: ISO/IEC 42001:2023 requirements: every clause explained →
What an ISO/IEC 27001 information security management system already gives an AIMS, what ISO/IEC 42001 adds, a mapping table of what to reuse and what to build, the risk and Statement of Applicability differences, and how a combined audit works.
Key points
- Moving from ISO 27001 to ISO 42001 is an extension, not a second project: both use ISO's harmonised structure, so the governance you already run can carry the AIMS.
- What must be added is AI-specific: an inventory of AI systems and your role for each, an AI policy, AI risk criteria, an impact assessment per system and a second Statement of Applicability against the 38 AI controls of ISO/IEC 42001 Annex A.
- The risk mechanism looks familiar but widens its lens: ISO/IEC 42001 requires consequences to individuals and societies, not only to the organisation and its information.
- A combined audit is possible only with a certification body accredited for both standards; check its ISO/IEC 42001 accreditation, under ISO/IEC 42006, separately.
Can an ISMS be extended into an AIMS?
Yes, and for most organisations that already hold ISO/IEC 27001 it is the sensible route. ISO/IEC 42001:2023 states in its introduction that it applies the harmonised structure, the same clause numbers, titles, core text and common terms that ISO/IEC 27001:2022 uses. Your ISMS therefore already has the machinery an AIMS needs: scope, policies, roles, a risk method, document control, audit, review and corrective action.
What the ISMS does not have is the subject matter. Information security asks whether information stays confidential, intact and available. AI governance asks whether each AI system is developed, provided and used responsibly, and what it could do to the people it touches. The extension adds that question to the machinery you own instead of building a parallel system.
What carries over from ISO/IEC 27001
In practice these usually need widening rather than writing:
- Context and interested parties (clause 4): the issues and interested-party registers gain AI entries; the climate-change question sits in clause 4.1 of both.
- Leadership and roles (clause 5): the role map gains an AIMS lead and an owner per AI system.
- Objectives and planning of changes (clause 6): one objectives register and one change procedure can hold both sets.
- Support (clause 7): document control, competence and awareness extend to AI roles.
- Performance evaluation and improvement (clauses 9 and 10): one audit programme, one review, one corrective action log.
Several information security controls also support AI systems directly: access management for models and training data, logging and monitoring, secure development, supplier security and incident management.
What ISO/IEC 42001 adds that the ISMS never asked for
Scope built on AI systems and roles
An ISMS scope names services, sites and systems. The AIMS scope names AI systems and whether you develop, provide or use each one, because the obligations differ by role. The scope guide covers how to draw the boundary.
An AI policy and AI objectives
Clause 5.2 asks for a separate AI policy that frames the AI objectives and refers, as relevant, to your other policies. It sits beside the security policy, not buried inside it.
The AI system impact assessment
Clause 6.1.4 has no ISO/IEC 27001 counterpart. For each AI system you assess what its deployment, intended use and foreseeable misuse could mean for individuals, groups and societies, and the result feeds the AI risk assessment. Clause 8.4 repeats it at planned intervals and on significant change. The impact assessment guide sets out the method.
A different Annex A, with Annex B guidance
ISO/IEC 42001 Annex A lists 38 reference controls in nine groups, from AI policies and resources to the system life cycle, data, information for interested parties, use of AI and third-party relationships. Annex B, also normative, gives implementation guidance for them. It is a different set from the 93 information security controls, not a copy of them. The Annex A guide explains each group.
Mapping table: ISMS element, what to reuse, what to add
| Element | Reuse from the ISMS | Add for the AIMS |
|---|---|---|
| Scope (4.3) | Sites, functions, interfaces with clients and suppliers | The AI systems in scope and your role for each; a separate documented AIMS scope |
| Policy (5.2) | Policy approval, communication and review routine | An AI policy that frames the AI objectives and references the security policy |
| Risk assessment (6.1.2, 8.2) | Method design, risk register format, review cycle | AI risk criteria; consequences for individuals and societies as well as the organisation |
| Risk treatment and SoA (6.1.3, 8.3) | Treatment workflow, plan format, residual-risk acceptance | Comparison with the 38 AI controls, Annex B guidance, a Statement of Applicability for ISO/IEC 42001 |
| Impact assessment (6.1.4, 8.4) | Nothing directly; privacy impact work, where it exists, helps | A documented assessment per AI system, repeated on change |
| Asset inventory and suppliers | Asset owners, supplier assessments, contracts | Data, tooling, computing and human resources per system; responsibilities with AI suppliers and customers |
| Competence and awareness (7.2, 7.3) | Competence matrix, training records | Competence for people who design, deploy and oversee AI; AI awareness for all staff |
| Audit, review, improvement (9, 10) | One programme, one meeting, one corrective action log | AI clauses and controls in the audit plan; AI inputs and decisions in the review |
Risk treatment and the Statement of Applicability: similar mechanism, different lens
The treatment clause reads almost the same in both standards: select options, determine the controls, compare them with Annex A so nothing necessary is missed, produce a Statement of Applicability and a treatment plan, and approve the residual risk. Three differences catch ISMS teams out.
- Who is harmed. The AI risk assessment analyses consequences for the organisation, for individuals and for societies. A register copied from the ISMS, scored only on business impact, will not meet it.
- Who approves. ISO/IEC 27001 requires risk owners to approve the treatment plan; ISO/IEC 42001 requires the designated management. Name them in the method so the auditor can see who signed.
- Where the controls come from. The AI Annex A is not exhaustive, so an information security control can appear in the AI treatment plan when it is genuinely needed.
Keep the two Statements of Applicability distinct, each against its own Annex A,. The risk assessment and treatment guide covers the AI method in depth.
A practical sequence for the extension
- Inventory and scope. List every AI system with an owner and your role; agree which the certificate covers.
- Gap analysis against the AI content. Sterling's gap analysis measures the existing ISMS against clauses 4 to 10 and the 38 AI controls, marking what can be kept as it is.
- Policy, criteria and assessments. Approve the AI policy, extend the risk method, then assess each system.
- Treatment and evidence. Produce the AI Statement of Applicability and plan, put life-cycle, data and supplier records into operation, and train the people involved.
- One audit, one review. Add AI to the audit programme and review agenda, and close findings before the certification body arrives.
Organisations extending an ISO/IEC 27001 system usually sit at the shorter end of the typical four-to-eight-month range; the number of AI systems sets the pace. Project phases and owners are in the implementation plan guide.
How Sterling extends an ISMS into an AIMS
One management system for security and AI, audited as one.
The Sterling Business Excellence Framework™ brings any combination of management system standards together as one system with one audit. Our expert and experienced consultants keep the ISMS processes that work, add the AI inventory, policy, impact and risk assessments with your system owners, and hold both standards in Effivity, included with the engagement at no additional cost; you choose Design™, where we guide your team, or Drive™, where we manage the work. We ensure successful certification for our clients. An independent, accredited certification body audits and decides; Sterling does not certify.
Book a free discovery callThe combined audit and the certification body
Certification bodies audit information security management systems under ISO/IEC 17021-1 with ISO/IEC 27006-1, and AI management systems under ISO/IEC 17021-1 with ISO/IEC 42006:2025, published in July 2025. Accreditation is granted per standard, and accredited ISO/IEC 42001 certification is still new, so your current body may not hold it. Ask which accreditation body has accredited it for ISO/IEC 42001 and confirm that on the accreditation body's public register.
Where one body holds both accreditations, the ISO/IEC 42001 audit can be combined with an ISO/IEC 27001 audit. The accreditation rules for combined audits calculate the audit time for each standard, then allow a justified reduction for the degree of integration, which a genuinely shared system earns more of than two stapled together. Choosing the body is covered in the certification body guide, and what the certificate does not prove under European law in the EU AI Act guide.
ISO/IEC 27001 to ISO/IEC 42001: questions buyers ask
Do our ISO 27001 Annex A controls count towards ISO 42001?
Not directly: ISO/IEC 42001 has its own Annex A of 38 AI controls, so each must be considered in a separate Statement of Applicability. Security controls such as access, logging and supplier assessment still support AI systems and can be named in the AI treatment plan where they are needed.
Can the ISO 42001 scope be narrower than our ISO 27001 scope?
Yes, each certificate carries its own scope, and the AIMS scope is defined by AI systems and roles rather than by services and sites. Starting with the systems customers ask about is common; widening later means extra audit time.
Can our ISO 42001 audit run with our ISO 27001 surveillance audit?
It can, if one certification body is accredited for both standards and the audit programme is planned for it. The body then aligns the two certificate cycles.
Does our ISO 27001 risk register work for ISO 42001?
Its format and review cycle do, but its content does not: the AI risk assessment must weigh consequences for individuals and societies against AI risk criteria.
Who should own the AIMS if the CISO owns the ISMS?
Usually a named AIMS lead with authority over the AI systems, working inside the same governance as the CISO. Impact assessment, data and responsible use also need product, data science and legal owners.
Does Effivity hold ISO 27001 and ISO 42001 in one place?
Yes, one Effivity environment holds documents, risk registers, treatment plans, audits, corrective actions and management reviews for both, with the AI system inventory set up as a configured register. The Effivity guide explains the set-up.
More ISO/IEC 42001 guides
Every guide in the ISO/IEC 42001 series is written for one question and links back to the hub for the whole picture.
- AI system impact assessment, with ISO/IEC 42005
- ISO/IEC 42001 Annex A: the 38 controls explained
- ISO/IEC 42001 certification process, step by step
- ISO/IEC 42001 certification cost: what moves the price
- Do I need ISO 42001? A decision guide to ISO/IEC 42001 for AI providers and users
- ISO 42001 documentation
- Effivity for ISO/IEC 42001: the guide
- ISO/IEC 27001 to ISO/IEC 42001: extending an ISMS into an AIMS
- ISO 42001 gap analysis
- ISO 42001 for healthcare
- ISO 42001 for HR and staffing
- ISO/IEC 42001 requirements explained, clause by clause
- ISO/IEC 42001 and the EU AI Act: what the certificate does not give
Keep exploring
Explore more about ISO/IEC 42001
Watch a video on how to implement ISO 42001 here.
Watch a video on ISO 42001 requirements here.
ISO/IEC 42001 resources
Services for ISO/IEC 42001
Talk to a consultant
Extend the ISMS you already run into an AIMS, without building a second system.
Tell us your goals, locations and target date. We’ll help you define the scope, timeline and next steps.
hello@sterling-grc.com+1 201 263 4443 (global headquarters) · your nearest office and its number are at the top of this page
What would you like to achieve?
Choose a starting point. The enquiry form takes about a minute.
Book my free discovery callContinue to the enquiry form to share your details. Please do not include passwords, sensitive personal information or confidential files. A Sterling consultant replies within one working day. Privacy notice