ISO 42001 Annex A is a set of 38 controls grouped under nine objectives, from AI policies and internal organisation to the AI system life cycle, data and third party relationships. They are not a mandatory checklist. An organisation selects the controls its AI risk and impact assessment calls for and records each inclusion or exclusion, with reasons, in a Statement of Applicability.
Teams approaching ISO 42001 often want to jump straight to the controls, the way they do with other standards. The controls matter, but Annex A only makes sense once you know it is a reference set selected by risk, not a list to implement wholesale. This explains what the 38 controls cover, how they are chosen, and how they differ from ISO 27001.
What is ISO 42001 Annex A?
Annex A is the normative set of controls for an AI management system. It holds 38 controls grouped under nine control objectives, numbered A.2 to A.10. The objectives move from governance at the top (policies, internal organisation, resources) through the AI system itself (impact assessment, life cycle, data) to how the organisation informs others and works with third parties.
Normative means Annex A is part of what an organisation is audited against, but not that every control applies. Which controls apply is decided by the AI risk and impact assessment, and the choice is documented in a Statement of Applicability. Annex B gives implementation guidance for each control, and Annex C lists potential AI related objectives and risk sources to weigh during the assessment, but only Annex A is normative.
What are the nine Annex A control objectives?
| Objective | Covers |
|---|---|
| A.2 Policies related to AI | Management direction and AI policy |
| A.3 Internal organisation | Roles, responsibilities and reporting for AI |
| A.4 Resources for AI systems | Data, tooling, compute and human resources |
| A.5 Assessing impacts of AI systems | The AI system impact assessment on individuals and society |
| A.6 AI system life cycle | Responsible design, development and deployment |
| A.7 Data for AI systems | Data quality, provenance and handling for AI |
| A.8 Information for interested parties | Transparency and documentation for users and affected parties |
| A.9 Use of AI systems | Responsible, intended use and monitoring in operation |
| A.10 Third party and customer relationships | Allocating responsibilities across suppliers and customers |
The objective that has no counterpart in information security is A.5. Assessing the impact of an AI system on people, covering fairness, bias, transparency and safety, is the heart of what ISO 42001 adds, and it drives which other controls an organisation needs.
How are the controls selected?
By risk, not by default. The organisation runs an AI risk and impact assessment, decides which Annex A controls the risks and interested party requirements call for, and records each inclusion and exclusion, with justification, in the Statement of Applicability. An auditor reads the Statement of Applicability first, so vague or unjustified decisions surface quickly. This is the same selection model as ISO 27001, which is one reason the two integrate cleanly.
How does ISO 42001 Annex A differ from ISO 27001 Annex A?
Different count, different focus. ISO 27001 Annex A has 93 information security controls in four themes (organisational, people, physical, technological). ISO 42001 Annex A has 38 controls in nine objectives focused on governing AI responsibly, including the impact assessment that information security does not address. An organisation running both selects from each set against its own risk assessment and keeps a Statement of Applicability for each. To gauge where to start, the ISO 42001 readiness checklist sets out what to have in place. We run ISO 42001 readiness for Australian organisations.
Frequently asked questions
How many controls does ISO 42001 Annex A have?
ISO 42001 Annex A is a normative set of 38 controls grouped under nine control objectives, numbered A.2 to A.10. They cover AI policies, internal organisation, resources, impact assessment, the AI system life cycle, data, information for interested parties, use of AI systems, and third party relationships.
Are all 38 ISO 42001 Annex A controls mandatory?
No. Annex A is a reference set, not a checklist. An organisation selects the controls its AI risk and impact assessment calls for and records each inclusion or exclusion, with justification, in a Statement of Applicability, the same model as ISO 27001.
What is the difference between ISO 42001 Annex A, Annex B and Annex C?
Annex A lists the controls. Annex B gives implementation guidance for them. Annex C lists potential AI related organisational objectives and risk sources to consider during the risk and impact assessment. Only Annex A is normative; B and C are guidance.
How does ISO 42001 Annex A differ from ISO 27001 Annex A?
ISO 27001 Annex A has 93 information security controls in four themes. ISO 42001 Annex A has 38 controls in nine objectives focused on governing AI, including assessing the impact of AI systems on people, which has no ISO 27001 equivalent.
What is the AI system impact assessment in ISO 42001?
A distinct requirement to assess the potential consequences of an AI system for individuals, groups and society, covering fairness, bias, transparency, safety and accountability. It is broader than information security risk and drives which Annex A controls apply.
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.
Sources:
- ISO/IEC 42001:2023, Artificial intelligence management system, iso.org, 2023
- ISO/IEC 27001:2022, Information security management systems, iso.org, 2022
Last updated: 19 October, 2026
