Guide

ISO/IEC 42001 for Australian organisations: what it asks and whether to certify

By Brett Raven, APAC Lead & Board Member, AIREUpdated 30 September 2026

ISO/IEC 42001 is the international standard for an AI management system: the policy, roles, risk and impact assessments, controls and reviews an organisation uses to run AI responsibly. ISO published it in December 2023, and Standards Australia adopted it unchanged as AS ISO/IEC 42001:2023 in February 2024. It applies to any organisation that develops, provides or uses AI, including one that only uses tools it buys. Certification is voluntary. The government's approach to AI rests on existing laws and voluntary guidance, so the usual reason to certify is a customer or a government contract that asks for it. Most organisations should build the basics first (an AI policy, named owners, a list of where AI is used and an assessment of its risks) and certify when someone they sell to wants the certificate.

What ISO/IEC 42001 is

General information, not compliance or legal advice. Read the standard itself, and take advice on your own obligations, before you commit to certification.

ISO/IEC 42001:2023 is titled Information technology, Artificial intelligence, Management system. ISO published the first edition in December 2023 and calls it the world's first AI management system standard. It specifies requirements for establishing, implementing, maintaining and continually improving an AI management system.

ISO defines that system as the policies, objectives and processes an organisation uses for the responsible development, provision or use of AI. Use is the word that matters. ISO says the standard is for organisations of any size that develop, provide or use AI-based products or services, in any industry, public sector agencies and not-for-profits included. A firm that has only switched on Copilot is in scope.

It is a management system standard, in the same family as ISO 9001 and ISO/IEC 27001. ISO sets it apart from its other AI standards on that point: it does not look at the detail of particular AI applications. It gives the organisation a way to manage AI risks and opportunities across the board, run on the Plan-Do-Check-Act cycle.

The Australian Standard

Standards Australia adopted it in February 2024 as AS ISO/IEC 42001:2023. It is an identical adoption: the Australian Standard and the ISO standard are the same text, so a reference to either means the same requirements.

What the standard asks for

The standard has two halves. The clauses set out the management system itself, from the AI policy to internal audit. Annex A then gives a reference set of controls to choose from.

What ISO/IEC 42001 covers, as Standards Australia summarises it
Part	What it asks for
Leadership and policy	An AI policy, and roles and accountabilities across the AI life cycle.
Planning	AI risk assessment and treatment, a process for AI system impact assessment, and measurable AI objectives.
Support	Resources (data, tooling, compute and people), competence, awareness, communication and documented information.
Operation	Operational planning and control, ongoing risk assessment and treatment, and periodic impact assessments.
Performance evaluation	Monitoring and measurement, internal audit and management review.
Improvement	Continual improvement, and dealing with nonconformity through corrective action.
Annex A (normative)	Reference control objectives and controls. Standards Australia's examples: AI policy, roles, data quality, information for users, logs, human oversight, and suppliers and customers.
Annex B (normative)	Implementation guidance for every Annex A control, including data quality, provenance, acquisition and preparation.
Annexes C and D (informative)	Example objectives and risk sources, and how to run the standard alongside other management systems such as ISO/IEC 27001, ISO/IEC 27701 and ISO 9001.

From Standards Australia's spotlight on AS ISO/IEC 42001:2023. The standard's own text is sold by ISO and Standards Australia. See the sources.

What you end up holding

Standards Australia lists the documents an organisation running the standard typically keeps: an AI policy, the scope of the management system, an AI risk assessment and treatment plan with a statement of applicability, AI system impact assessments, competence records, monitoring and event logs, internal audit reports and management review outputs.

Two things on that list catch organisations out. The impact assessment is about the effect of each AI system, which is a different question from the risk to the organisation. And internal audit and management review mean the system has to run, on a schedule, with records. A policy on the intranet is a start. It is not a management system.

How it fits with ISO 27001 and privacy law

ISO/IEC 27001

ISO/IEC 27001 sets the requirements for an information security management system. ISO/IEC 42001 does the same job for AI. Standards Australia notes that 42001 shares ISO's harmonised structure, and that its Annex D covers running it alongside ISO/IEC 27001 and ISO/IEC 27701 so the AI controls sit next to the security and privacy ones. BSI describes 42001 as extending the principles of 27001 to the security and data protection problems specific to AI.

If you already hold ISO/IEC 27001, you have built the management machinery once: a scope, a risk method, internal audit and management review that people turn up to. The AI work is new. Impact assessments, data quality and provenance, human oversight and the information you give users need work a security programme will not have done. Neither certificate covers the other.

ISO/IEC 23894

ISO/IEC 23894 gives guidance on managing AI risk. ISO sets it beside 42001 and calls 42001 the management system standard, the one with requirements. BSI names 23894 as a companion that helps with implementation. Use it for the method of your AI risk assessment. You certify against 42001.

The Privacy Act

A certificate changes nothing about your privacy obligations. The Office of the Australian Information Commissioner says the Privacy Act applies to every use of AI that involves personal information, and that the obligations cover what goes into an AI system and what comes out of it when the output contains personal information. As a matter of best practice, it recommends that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools. It also recommends a privacy by design approach, including a privacy impact assessment.

One date to put in the plan. From 10 December 2026, APP entities that use personal information in automated decision-making with the potential to affect people's rights or interests must say in their privacy policies what kinds of personal information are used and what kinds of decisions are made that way. The OAIC consulted on its guidance for this obligation in May and June 2026.

Run the privacy impact assessment and the standard's AI system impact assessment together. They overlap, and neither replaces the other.

The Australian policy context

We know of no Australian law that requires ISO/IEC 42001 as at September 2026. The government's National AI Plan, released on 2 December 2025, says its regulatory approach to AI will build on existing legal and regulatory frameworks, with agencies and regulators responsible for AI harms in their own domains. The plan announced an AI Safety Institute to monitor and test emerging AI and advise regulators. Beyond the law, it promotes responsible practice through voluntary measures and guidance.

From the Voluntary AI Safety Standard to the Guidance for AI Adoption

The Department of Industry, Science and Resources published the Voluntary AI Safety Standard in September 2024: ten voluntary guardrails, which the department says align with international standards including ISO/IEC 42001:2023 and the US NIST AI Risk Management Framework. Being voluntary, it created no new legal duties.

On 21 October 2025 the department published the Guidance for AI Adoption, which it describes as evolving the Voluntary AI Safety Standard. The National AI Centre condensed the ten guardrails into six essential practices and extended the guidance to AI developers. It comes in two versions: foundations, for organisations early in their AI use, and implementation practices, for governance professionals and technical experts. The implementation practices say they draw on and align with AS ISO/IEC 42001:2023, and that every Voluntary AI Safety Standard practice is now inside them. The six practices are:

Decide who is accountable.
Understand impacts and plan accordingly.
Measure and manage risks.
Share essential information.
Test and monitor.
Maintain human control.

An organisation that works through the implementation practices has done much of what ISO/IEC 42001 asks for. What it will not have is an audited management system or a certificate.

If you are a Commonwealth agency

The Digital Transformation Agency's Policy for the responsible use of AI in government, version 2.0, took effect on 15 December 2025. It applies to non-corporate Commonwealth entities, with some exceptions. Its mandatory requirements include accountable officials, transparency statements, a strategic approach to AI adoption, accountability for each AI use case, internal use case registers, staff training on AI and an impact assessment for each AI use case. None of it requires certification. Much of it lines up with the standard's roles, competence and impact assessment requirements.

If you sell to government

The Commonwealth published model clauses for buying AI in March 2025, developed by the Digital Transformation Agency with the Australian Government Solicitor. According to the Solicitor's summary, a buyer can require the seller to set up an AI risk management system through either a short clause requiring compliance with ISO/IEC 42001:2023 or a detailed clause setting out what the system must cover. The short clause asks for compliance, not a certificate. What evidence satisfies the buyer is a question for the contract, and a certificate from an accredited body is the easiest evidence to hand over.

Do you need certification, or just alignment?

Alignment means running your AI governance to the standard without the external audit. Certification means an accredited body audits it and issues a certificate. Both need the same work inside the organisation. Certification adds the external audits and their fees.

Certification is worth it when

A customer asks for the certificate by name, in due diligence, a security questionnaire or a tender.
You build or sell AI-based products or services, and your buyers need independent evidence of how you manage them.
A government contract includes the ISO/IEC 42001 clause and the buyer wants more than your word for it.
You already hold ISO/IEC 27001, so the management system exists and the extra audit is a smaller step.

Alignment is enough when

You use AI tools other companies build, and nobody you sell to has asked for the certificate.
Your main risk is staff pasting the wrong thing into a prompt. A written rule and the right settings on each tool deal with most of that.
You want the discipline of the standard and can wait to see whether the market asks for proof.

Either way, nothing is wasted. The work that makes an organisation ready for certification is the same work that makes its AI use defensible without it.

A practical path to readiness

Work in this order. Each step feeds the next, and the first three are worth doing whether or not you ever certify.

List where AI is used. Every tool staff use, including on personal accounts, every AI feature switched on inside software you already pay for, and every model inside a product you sell. You cannot govern what you have not found.
Set the AI policy. Which tools are approved and for what, what may go into a prompt and what may not, who checks output before it reaches a customer or commits money, and who to ask when it is unclear. Our one-page AI policy template is a starting point.
Name the owners. One person accountable for the management system, and an owner for each AI system. The first practice in the Guidance for AI Adoption is the same: decide who is accountable.
Assess risks and impacts. For each system, the risk to the organisation and the impact on the people it affects. Run the privacy impact assessment at the same time.
Choose and apply controls. Use Annex A as the reference list, record what you chose in a statement of applicability, and make the controls real in each tool's settings and each supplier's contract.
Run it, then decide. Monitor, hold an internal audit and a management review, fix what they find. Decide on certification once the system has run for long enough to show an auditor it works.

Do not skip the last step. A Stage 2 auditor checks that the requirements are implemented and followed, which means records of the system running. A folder of documents written the month before the audit does not show that.

What certification involves, and who does it

ISO writes the standard. Certificates come from certification bodies, and Standards Australia notes that certification arrangements sit outside the standard's text. What makes a certificate worth having is accreditation: as ISO puts it for ISO/IEC 27001, an accreditation body has independently confirmed the certification body's competence.

Accreditation in Australia

JASANZ is the joint accreditation body for Australia and New Zealand, established by a treaty between the two governments. It says it does not certify products, management systems or people: it accredits the organisations that do. Its Business & Innovation sector lists an Artificial Intelligence Management System Scheme, and Intertek announced in 2025 that JAS-ANZ had accredited it to certify organisations to ISO/IEC 42001:2023. Some certification bodies working in Australia hold their accreditation from overseas instead. BSI, for one, is accredited for the standard by UKAS in the UK, RvA in the Netherlands and ANAB in the US.

ISO/IEC 42006:2025, published in July 2025, sets the extra requirements for bodies that audit and certify AI management systems, building on ISO/IEC 17021-1. Ask any body you are considering which accreditation covers its ISO/IEC 42001 certificates, and check the certificate on the accreditation body's register. JASANZ publishes registers of accredited bodies and certified organisations.

The audit cycle

Intertek SAI Global describes the path for its management system certifications. It starts with a pre-assessment and self-evaluation, a gap analysis before the first audit. A Stage 1 audit follows, where the auditor reviews your documents and readiness. The Stage 2 audit is a full assessment of whether the requirements are implemented and followed. After certification come surveillance audits, generally yearly, and recertification, generally every three years.

We do not quote a timeline or a price for any of this. Both depend on the scope you choose and how much of the system already runs, and the certification body sets its own fees and audit dates.

Keep the adviser and the auditor apart

The certification bodies we checked keep consulting away from what they certify. BSI says its certification arm will not certify a management system that another part of BSI consulted on, and Intertek says its assurance business does not provide consulting for management system certification. Expect the firm that helps you build the system and the firm that audits it to be different firms.

How AIRE helps

We write AI governance that people follow, and much of it maps to the standard. Our AI strategy consulting and governance work produces a one-page data-handling rule, which is a data control. The AI policy is the policy the standard names, and it says who owns AI risk. The position the board signs is the leadership commitment the standard asks for.

It covers several of the standard's requirements, not all of them. An audit also wants internal audit and management review running over time, and that is yours to run. We do not certify, and we do not claim certification. That takes an accredited certification body.

If you are not sure where you stand, start with the free AI readiness assessment. It takes about three minutes and gives you a score by category and where we would start.

Asked for ISO/IEC 42001 in a tender? The discovery call is free and takes 30 minutes.

Frequently asked questions

Is ISO 42001 mandatory in Australia?

No. ISO/IEC 42001 is a voluntary standard, and certification to it is a choice. The Australian Government's National AI Plan (December 2025) builds its approach to AI on existing laws and regulators, with the National AI Centre's Guidance for AI Adoption as voluntary guidance. It can become a requirement through a contract: the Commonwealth's AI model clauses give government buyers the option of requiring a seller to comply with it. Laws that already apply to your organisation, such as the Privacy Act, apply to its use of AI whether or not you certify. This is general information, not legal advice.

Is ISO 42001 the same as ISO 27001?

No. ISO/IEC 27001 sets the requirements for an information security management system. ISO/IEC 42001 sets them for an AI management system: an AI policy, AI risk and impact assessment, and controls such as data quality, logs and human oversight. The two share ISO's harmonised structure, so if you already run ISO/IEC 27001 the machinery of internal audit and management review will be familiar, and Annex D of ISO/IEC 42001 covers running them together. A 27001 certificate does not cover AI management.

Who can certify us to ISO 42001?

A certification body accredited for ISO/IEC 42001. In Australia and New Zealand the accreditation body is JASANZ, set up by treaty between the two governments, and it accredits the certification bodies rather than certifying anyone itself. Intertek, for example, announced JAS-ANZ accreditation for ISO/IEC 42001 in 2025, and BSI holds accreditation from UKAS, the Netherlands' RvA and the US's ANAB. Ask which accreditation body stands behind a certificate before you sign, and check it on that body's register.

What does an ISO 42001 certification audit involve?

Intertek SAI Global describes the path for its management system certifications like this. A pre-assessment and self-evaluation first. Then a Stage 1 audit, where the auditor reviews your documents and readiness, and a Stage 2 audit, a full assessment of whether the requirements are implemented and followed. After certification, surveillance audits are generally yearly and recertification generally every three years. Confirm the terms with the body you choose.

Is there an Australian version of ISO 42001?

Yes, and it is the same text. Standards Australia adopted ISO/IEC 42001:2023 as an identical Australian Standard, AS ISO/IEC 42001:2023, in February 2024. You can buy it from Standards Australia or from ISO.

How does ISO 42001 relate to the Guidance for AI Adoption?

They point the same way. The National AI Centre's Guidance for AI Adoption, published in October 2025, is the first update of the 2024 Voluntary AI Safety Standard and condenses its ten guardrails into six essential practices. Its implementation practices say they draw on and align with AS ISO/IEC 42001:2023. The guidance is voluntary and has no certificate. ISO/IEC 42001 is the one an auditor can certify you against.

Does AIRE certify organisations to ISO 42001?

No. We do not certify, and we do not claim certification. That takes an accredited certification body. Our governance work covers several of the standard's requirements: an AI policy, a data-handling rule, named owners and the leadership position the standard asks for. It does not provide the internal audit, the management review or the certification audit.

Sources

Every fact on this page about the standard, certification and Australian policy comes from the sources below. We read each one on 30 September 2026. The standard's own text is sold by ISO and Standards Australia, so its contents are described here as Standards Australia and ISO summarise them. Policy and guidance change, so check anything that matters before you act on it.

ISO, ISO/IEC 42001:2023, AI management systems (published December 2023, edition 1). Read 30 September 2026.
Standards Australia, Standards Australia adopts the international standard for AI Management System, AS ISO/IEC 42001:2023 (16 February 2024). Read 30 September 2026.
Standards Australia, Spotlight on: AS ISO/IEC 42001:2023, Artificial intelligence, Management system (5 September 2025). Read 30 September 2026.
ISO, ISO/IEC 27001:2022, Information security management systems. Read 30 September 2026.
ISO, ISO/IEC 42006:2025, Requirements for AIMS audit and certification bodies (published July 2025). Read 30 September 2026.
JASANZ, Australia & New Zealand's Accreditation Authority (home page, with links to its registers of accredited bodies and certified organisations). Read 30 September 2026.
JASANZ, What is accreditation?. Read 30 September 2026.
JASANZ, Leadership & Governance. Read 30 September 2026.
JASANZ, Business & Innovation (lists the Artificial Intelligence Management System Scheme). Read 30 September 2026.
Intertek, Intertek achieves global JAS-ANZ accreditation for ISO/IEC 42001:2023 Artificial Intelligence Management System. Read 30 September 2026.
Intertek SAI Global, 5 Steps to Certification (for its management system certifications in general). Read 30 September 2026.
BSI Australia, ISO/IEC 42001 AI Management System. Read 30 September 2026.
BSI Australia, ISO/IEC 42001: Why certify your AI Procedures and getting started with BSI. Read 30 September 2026.
Department of Industry, Science and Resources, Voluntary AI Safety Standard (published 5 September 2024, updated 2 December 2025). Read 30 September 2026.
Department of Industry, Science and Resources, Voluntary AI Safety Standard: the 10 guardrails. Read 30 September 2026.
National AI Centre, Essential AI practices. Read 30 September 2026.
National AI Centre, Guidance for AI adoption: implementation guidance. Read 30 September 2026.
National AI Centre, Guidance for AI adoption: implementation guidance (PDF, October 2025). Read 30 September 2026.
Department of Industry, Science and Resources, National AI Plan: Keep Australians safe (published 2 December 2025). Read 30 September 2026.
Digital Transformation Agency, Policy for the responsible use of AI in government, version 2.0 (last updated 1 December 2025). Read 30 September 2026.
Australian Government Solicitor, Legal update no. 327: Artificial Intelligence (AI) model clauses (4 April 2025). Read 30 September 2026.
Office of the Australian Information Commissioner, Guidance on privacy and the use of commercially available AI products (published 21 October 2024, updated 17 January 2025). Read 30 September 2026.
Office of the Australian Information Commissioner, Consultation on Guidance for Transparency in Automated Decision Making (published 18 May 2026). Read 30 September 2026.

A one-page AI policy template for Australian organisations
Copy a one-page AI policy for Australian organisations: approved tools, what data may go in, who checks output, who owns AI risk, and how to roll it out.
What AI consulting costs in Australia
What an AI consultant costs in Australia: AIRE's own prices, what federal AI advisory contracts on AusTender show, and how to compare quotes.

Related services

AI strategy consulting and governance: an AI policy, a one-page data-handling rule and a position the board can sign.
AI readiness assessment: a free, short assessment of where your organisation stands on AI.

All AIRE guides