ISO/IEC 42001:2023

ISO 42001:2023. Govern AI with a real management system.

CyberHeed builds your AI Management System end to end - AI policy, impact assessments, lifecycle controls, evidence and surveillance. The first AI standard, handled like the management system it is.

38 Annex A Controls
9 Control Topic Areas
3 yr Certification Cycle
First management system standard for AI
Issued December 2023
Harmonized Structure with ISO 27001
Certifiable by accredited bodies
What ISO 42001 Is

A management system for AI. Not a checklist of AI principles.

ISO/IEC 42001:2023 is the first international management system standard for Artificial Intelligence. Issued in December 2023 by ISO/IEC JTC 1 SC 42, it specifies the requirements for establishing, implementing, maintaining and continually improving an AI Management System (AIMS). It applies to any organisation that provides or uses products or services that utilise AI systems, regardless of size or industry.

It addresses what classical IT management does not: automated decision-making that can be non-transparent, data-driven systems whose behaviour comes from training rather than human-coded logic, and continuous learning systems that change in production.

Who Needs ISO 42001?

ISO 42001 applies to any organisation that develops, deploys, procures or uses AI systems. It is particularly relevant for:

AI Developers and Providers

Companies building AI products, machine learning platforms, foundation models, or AI-powered features. ISO 42001 demonstrates responsible AI practice to customers, partners and regulators - and is widely discussed as a supporting standard for conformity assessment of high-risk AI systems under the EU AI Act.

AI Deployers and Integrators

Organisations embedding third-party AI into their products and services, or using AI to make decisions about customers, employees or operations. ISO 42001 establishes accountability across the AI supply chain and meets growing buyer expectations on responsible AI.

Regulated Sectors

Health, finance, transport, employment, energy and defence. Annex D of the standard explicitly names these as sectors where AI Management Systems are particularly applicable. Sector regulators are increasingly looking to ISO 42001 as the reference for AI governance.

Public Sector and Government Suppliers

AI procurement is rapidly being formalised in government and public sector tenders. Organisations seeking to supply AI capability to government will increasingly need to demonstrate a credible AI management system.

The Certification Journey

Like other ISO management system standards, ISO 42001 certification follows a structured cycle delivered by accredited certification bodies:

1. Gap Assessment

Identify where current AI practices fall short of ISO 42001 requirements. Map AI systems, organisational roles, and existing governance against the clauses and Annex A controls.

2. AIMS Implementation

Build the AI Management System: AI policy, risk assessment, AI system impact assessment, Statement of Applicability, lifecycle controls, data governance, human oversight, and training.

3. Internal Audit

Conduct internal audits to verify the AIMS meets all requirements before external assessment. Close nonconformities and confirm impact assessments are complete and current.

4. Stage 1 Audit (Documentation)

The certification body reviews documentation: AI policy, Statement of Applicability, AI risk assessment, AI system impact assessments, and management system records. They confirm readiness for Stage 2.

5. Stage 2 Audit (Implementation)

The auditor verifies the AIMS is implemented and operating effectively across the AI system lifecycle. They interview staff, sample controls, and test how AI risks and impacts are actually being managed.

6. Year 1 & 2 Surveillance

Annual surveillance audits verify continued compliance. AI moves fast - models change, data shifts, new systems are deployed. Evidence must be current, and the AIMS must keep pace.

7. Year 3 Recertification

Full recertification audit. The entire AIMS is reassessed. The cycle begins again for another three years.

The Management System

Clauses 4 to 10: the same Harmonized Structure as ISO 27001

ISO 42001 uses the Harmonized Structure shared with ISO 27001, ISO 9001 and other ISO management system standards. The clauses establish how AI is governed, risk-assessed, resourced, operated and improved - so that controls are part of a system, not a one-off project.

Clause 4 - Context of the Organisation

Understand internal and external context, the needs of interested parties, and define the scope of the AIMS. Critically, the organisation must determine its role(s) with respect to AI - provider, producer, customer, partner, subject, or relevant authority - because this determines which requirements and controls apply.

Clause 5 - Leadership

Top management must demonstrate commitment and establish an AI policy aligned with the organisation's strategic direction. The note to clause 5.1 explicitly highlights modelling a responsible AI culture as a demonstration of leadership.

Clause 6 - Planning

Identify AI risks and opportunities. Conduct AI risk assessment, AI risk treatment, and - distinctly - AI system impact assessment covering individuals, groups and societies. Set measurable AI objectives. Produce a Statement of Applicability covering Annex A controls.

Clause 7 - Support

Provide resources, ensure competence (AI demands interdisciplinary skills), build awareness, manage communication and documented information. People, capability and infrastructure for the AIMS to operate.

Clause 8 - Operation

Execute the planned AI risk assessments, risk treatments and impact assessments. Operate the AI system lifecycle controls (Annex A.6). Retain documented evidence that each was carried out as planned.

Clause 9 - Performance Evaluation

Monitor, measure, analyse and evaluate the AIMS and AI system performance. Conduct internal audits. Perform management reviews. AI systems drift - this clause ensures the AIMS sees it.

Clause 10 - Improvement

Address nonconformities with corrective actions. Drive continual improvement of the AIMS. AI moves faster than most management systems - the AIMS has to be built to keep up.

Annex A Controls

38 controls across nine topic areas. Every one mapped in CyberHeed.

Annex A of ISO 42001 contains 38 reference controls organised into nine topic areas (A.2 to A.10). As with ISO 27001, every control must be considered in the Statement of Applicability, with justification for inclusion or exclusion based on the AI risk assessment.

A.2 Policies Related to AI (3 controls)

AI policy, alignment with other organisational policies, and review of the AI policy. The governance anchor for everything that follows - documented direction on how AI is developed and used.

A.3 Internal Organisation (2 controls)

AI roles and responsibilities, and a process for reporting concerns about AI systems with confidentiality, anonymity and protection against reprisal. Accountability and a safe channel for raising AI-specific concerns.

A.4 Resources for AI Systems (5 controls)

Documentation of the resources needed across the AI system lifecycle: data, tooling, system and computing resources, and human competence. The foundation for understanding what an AI system is actually made of.

A.5 Assessing Impacts of AI Systems (4 controls)

A formal AI System Impact Assessment process - distinct from risk assessment - covering impacts on individuals, groups and societies. Fairness, accountability, transparency, safety, privacy, accessibility, human rights, and societal effects must all be considered.

A.6 AI System Lifecycle (9 controls)

The largest topic area. Management guidance for responsible development, plus lifecycle controls covering requirements, design and development documentation, verification and validation, deployment, operation and monitoring, technical documentation and event logging. AI-specific threats - data poisoning, model stealing, model inversion - are explicitly in scope.

A.7 Data for AI Systems (5 controls)

Data management for development and enhancement, acquisition, quality, provenance and preparation. Data governance is treated as a first-class concern because in AI, data is the system.

A.8 Information for Interested Parties (4 controls)

System documentation and user information, external reporting channels, incident communication, and obligations to inform regulators, customers and authorities. Transparency made operational.

A.9 Use of AI Systems (3 controls)

Processes for responsible use, objectives for responsible use (fairness, accountability, transparency, explainability, reliability, safety, privacy, security, accessibility), and ensuring AI is used within its intended use and documentation. Where and how human oversight is required is determined here.

A.10 Third-Party and Customer Relationships (3 controls)

Allocating responsibilities across the AI supply chain, supplier processes, and consideration of customer expectations. AI accountability follows the system - it does not stop at the contract boundary.

ISO 42001 is not a paper exercise. The AI System Impact Assessment, lifecycle controls and continuous monitoring requirements mean the AIMS has to keep pace with how AI actually evolves - retrained models, new datasets, drifting behaviour, and shifting regulation.

Prepare - Comply - Manage

How CyberHeed handles each phase of your ISO 42001 journey

Every phase of ISO 42001 compliance - from initial preparation through ongoing surveillance - managed in one platform. The same agentic GRC approach as our ISO 27001 implementation, adapted for AI-specific requirements.

1. Prepare with SmartPrep

An adaptive, AI-guided journey covering every ISO 42001 topic area. AI policy, AI risk and impact assessment, lifecycle and data governance, transparency and human oversight, supplier and customer responsibilities - each session adapts as the conversation unfolds, follows up on gaps, and captures how your organisation actually develops or uses AI.

At the end, your complete documentation suite is generated from the knowledge gathered - including the AI System Impact Assessments that are unique to this standard.

2. Comply with Evidence Validation

Upload evidence for any of the 38 Annex A controls. AI reads each document, assesses whether it satisfies the requirement, and tells you specifically what's strong and what an auditor would flag. Scored 0-5 with actionable feedback.

AutoMatch reads your existing documentation and maps it to the right controls across ISO 42001, ISO 27001 and any other framework you maintain. Hours of cross-referencing, handled automatically.

3. Manage for Surveillance

Surveillance audits become routine. Impact assessments are kept current as systems change. Model performance and data drift are tracked. Recurring tasks have owners and deadlines. Gaps are flagged before auditors find them.

When Year 2 and Year 3 come around, your posture is current. When recertification arrives, you're already ahead.

Start your ISO 42001 journey.

From AI policy to certification. One platform. AI that does the heavy lifting.

Book a Demo