ISO / IEC 42001 : 2023

The world's first
AI Management
System standard

Published December 2023. ISO/IEC 42001 gives organizations a globally recognized, auditable framework for responsible AI — covering the entire lifecycle from planning through deployment and ongoing monitoring.

Published Dec 2023 EU AI Act aligned First mover advantage
Start your gap analysis View packages
Standard at a glance
Full name
ISO/IEC 42001:2023 — Artificial Intelligence Management System
Published
December 2023 · First edition · ISO + IEC joint standard
Applies to
Any org that develops, provides, or uses AI systems — any size, any sector
Structure
High-level structure (HLS) — integrates with ISO 9001, 27001, 37001
Companion standards
ISO/IEC 23894 (risk), 42005 (impact), TR 24027 (bias) supplement 42001
ISO certifications via RMC partnership
What is it

A management system standard built for the AI era

ISO/IEC 42001 is not an AI ethics checklist or a technical specification — it is a full management system standard in the same family as ISO 9001 (quality) and ISO 27001 (information security).

It provides organizations with a framework to establish, implement, maintain, and continually improve their approach to AI governance. Just as ISO 9001 doesn't tell you how to manufacture a product but how to manage the quality of your manufacturing — ISO 42001 doesn't tell you how to build an AI model, but how to govern the AI systems your organization deploys.

The standard is built on the ISO High-Level Structure (HLS), meaning it integrates cleanly with existing management systems your organization may already hold. If you're certified to ISO 9001, ISO 14001, or ISO 37001, adding ISO 42001 shares documentation structure, internal audit cycles, and management review — reducing the total compliance burden significantly.

Discuss your existing certifications
How it came to be
2016–2019
ISO/IEC JTC 1/SC 42 established
ISO creates a dedicated subcommittee for AI standards as AI adoption accelerates across industry.
2020
Work begins on AI management systems
ISO initiates development of a management system standard specifically for organizations using AI — parallel to the EU AI Act drafting process.
2022
Draft International Standard (DIS) published
Public consultation on the draft — 57 member countries participate in shaping the final standard.
Dec 2023
ISO/IEC 42001:2023 published
The first edition of the world's first AI Management System standard is officially published. Certification audits begin immediately.
2024–2026
Global adoption accelerates
EU AI Act enforcement timelines drive rapid uptake. Regulated industries — banking, healthcare, critical infrastructure — prioritize certification. DeepSynergy leads SE Asia rollout.
Why now

The window for first-mover advantage is closing

Early adopters win contracts, build trust, and set industry benchmarks. Organizations that wait face certification backlogs and regulatory pressure simultaneously.

2026
EU AI Act enforcement begins

High-risk AI systems must demonstrate governance compliance from August 2026. ISO 42001 is the internationally recognized framework for doing so. Organizations without certification face legal exposure.

60%
Faster audit prep with our AI tooling

DeepSynergy's orchestration platform automates first-pass gap analysis, document indexing, and evidence collection — meaning your certification timeline is significantly shorter than competitors using manual methods.

#1
Competitive tender advantage

Enterprise procurement teams in regulated industries are already asking suppliers to demonstrate AI governance. ISO 42001 certification becomes a tender requirement in banking, healthcare, and government contracting by 2027.

What it covers

The standard's key requirement areas

ISO/IEC 42001 is structured around ten clauses — the first three are introductory, clauses 4–10 are the auditable requirements. Here's what each covers and what evidence your auditor will look for.

CL 4
Context of the organization
Define the internal and external issues that affect your AI use. Identify stakeholders, their needs, and the scope of your AI management system. Map all AI systems within scope.
AI inventoryStakeholder mapping
CL 5
Leadership & commitment
Top management must establish AI policy, assign roles and responsibilities, and actively demonstrate commitment to responsible AI. No delegation to a single "AI officer" — leadership accountability is audited.
AI policyBoard oversight
CL 6
Planning & AI risk management
Identify AI-specific risks — bias, safety, privacy, transparency failures. Assess impact and likelihood. Set measurable AI objectives. This is where your AI risk register lives and where most organizations have the largest gaps.
Risk registerAI objectives
CL 7
Support — resources, competence & awareness
Ensure adequate resources (people, tools, budget) for AI governance. Staff competency in AI must be assessed and documented. Awareness programs must be evidenced — not just implemented once.
Competency recordsTraining logs
CL 8
Operation — AI system lifecycle
The largest clause — covers planning and controlling AI system development, deployment, and decommissioning. Includes data governance, testing, human oversight, and supplier AI accountability. Where most operational evidence is generated.
Data governanceHuman oversightSupplier mgmt
CL 9
Performance evaluation
Monitor, measure, analyze, and evaluate AI system performance and your management system's effectiveness. Internal audits and management reviews are required — not optional. KPIs must be tracked and reported.
Internal auditManagement review
CL 10
Improvement
React to nonconformities with documented corrective actions. Drive continual improvement of the AI management system — not just fixing problems but proactively raising the bar. Improvement plans are audited at each surveillance cycle.
Corrective actionsContinual improvement
A1–A9
Annex A — AI-specific controls
38 additional controls addressing AI-specific areas: explainability, fairness, data quality, adversarial robustness, model documentation, and AI incident response. Organizations select applicable controls based on risk — similar to ISO 27001's Annex A.
ExplainabilityFairness controlsIncident response
The certification journey

What to expect — phase by phase

Most organizations achieve ISO 42001 certification within 4–9 months, depending on size, complexity, and existing management system maturity. Here's exactly what happens at each stage.

01
Weeks 1–4
Gap analysis & AI inventory 2–4 weeks

We map every AI system your organization deploys — internally built, vendor-supplied, or embedded in purchased software. Each system is assessed against all ISO/IEC 42001 clause requirements. You receive a gap analysis report showing your current compliance posture, risk exposures, and a prioritized remediation roadmap scored by effort and impact.

Clause-by-clause gap analysis report
Complete AI systems inventory register
Risk exposure summary with priority scoring
Remediation roadmap with effort estimates
02
Weeks 5–16
AI management system design 6–12 weeks

We build the documented management system your auditor will review — every policy, procedure, work instruction, and evidence template needed to satisfy clauses 4–10 and selected Annex A controls. Everything is tailored to your organization's actual AI systems, not generic templates. Staff awareness training runs in parallel so your team understands what they're signing up for.

AI policy & governance framework
AI risk management procedure + register
Data governance & explainability procedures
Annex A control selection & Statement of Applicability
AI objectives, KPIs & monitoring plan
Staff awareness training delivery
03
Weeks 17–19
Internal audit & pre-certification review 2–3 weeks

Before inviting the certification body in, we conduct a full internal audit simulating Stage 2. We identify any remaining nonconformances or observations, support corrective actions, and ensure your evidence package is complete and auditor-ready. This is the most valuable step — it eliminates surprises during the external audit.

Internal audit report with all findings
Corrective action plan (CAP) & evidence
Stage 1 documentation review support
Pre-audit readiness confirmation
04
Week 20+
Certification audit — Stage 1 & Stage 2 2–5 days on-site

Your chosen certification body conducts the two-stage external audit. Stage 1 is a documentation review (typically remote, 1 day). Stage 2 is the on-site assessment where auditors interview staff and examine evidence. We accompany you throughout — briefing your team before each session, supporting real-time responses, and managing any nonconformance responses after the audit closes.

Pre-audit staff briefing & simulation
On-site / remote audit accompaniment
Nonconformance response drafting
ISO 42001 certificate issued ✓
05
Annual
Ongoing surveillance & continuous improvement Annual cycle

ISO 42001 certification is valid for three years, with annual surveillance audits in years 1 and 2 and a recertification audit in year 3. We manage your ongoing compliance — updating your management system as your AI portfolio evolves, preparing evidence packages for each surveillance cycle, and driving continual improvement so your AI governance matures over time.

Annual surveillance audit preparation
Management system updates as AI evolves
Continual improvement roadmap
Year 3 recertification support
The transformation

Before and after ISO 42001

What your AI governance looks like today — and what it looks like after certification.

Before — unstructured AI use
No documented AI policy — each team uses AI differently with no governance
AI systems deployed without formal risk assessment — bias and safety risks unknown
No AI inventory — leadership doesn't know which AI systems the org actually uses
Data used to train models not governed — quality, privacy, and consent unclear
No audit trail — if an AI decision is challenged, there's no evidence of how it was made
Regulators and clients ask AI governance questions — no structured answer available
Tenders increasingly require AI governance proof — bids lost to certified competitors
After — ISO 42001 certified
Documented AI policy endorsed by leadership — consistent governance across the organization
Every AI system assessed against risk framework — bias, safety, and privacy risks documented and mitigated
Complete AI inventory maintained — leadership has full visibility of AI exposure
Data governance procedures in place — training data quality, privacy, and lineage tracked
Full audit trail for AI decisions — evidence package ready for regulatory review at any time
Structured responses to regulator and client AI questions — backed by certified evidence
ISO 42001 certificate unlocks tenders — win business that uncertified competitors cannot
EU AI Act · 2024

ISO 42001 is your EU AI Act compliance foundation

The EU AI Act came into force in August 2024, with high-risk AI system requirements applying from August 2026. ISO/IEC 42001 directly addresses the governance, risk management, and transparency requirements that high-risk AI operators must demonstrate. Certification now means compliance readiness when enforcement begins.

Risk management system (Art. 9)
ISO 42001 Clause 6 risk management satisfies EU AI Act Article 9 requirements for high-risk AI
Data governance (Art. 10)
ISO 42001 Clause 8 data governance procedures address EU AI Act Article 10 data quality requirements
Transparency & logging (Art. 12–13)
Annex A explainability and logging controls directly address transparency and traceability obligations
Human oversight (Art. 14)
ISO 42001 Clause 8 human oversight requirements map directly to EU AI Act Article 14 obligations
Aug
2026
High-risk AI system obligations apply — enforcement begins
85+
Countries with active AI regulation or regulatory roadmap
€30M
Maximum EU AI Act fine for prohibited AI practices (or 6% global turnover)
Discuss EU AI Act readiness
Common questions

Frequently asked about ISO 42001

How long does ISO 42001 certification take?
Most organizations achieve certification within 4–9 months from the start of the gap analysis. Organizations with existing ISO 9001 or ISO 27001 certifications tend to move faster (3–5 months) because the management system infrastructure is already in place. Organizations starting from zero with complex AI portfolios may take up to 12 months. DeepSynergy's AI tooling compresses the documentation and evidence-gathering phases significantly — our clients consistently achieve certification 30–40% faster than industry average.
Is ISO 42001 mandatory or voluntary?
ISO 42001 is currently voluntary — it is not directly mandated by any law. However, it is increasingly referenced as a compliance pathway in regulatory frameworks. The EU AI Act does not require ISO 42001 specifically, but certified organizations demonstrate risk management and governance practices that directly satisfy regulatory requirements. In practice, ISO 42001 is becoming a de facto requirement in regulated industry supply chains — banks, insurers, and government agencies are beginning to require it of AI-using suppliers. Voluntary today, industry-required by 2027 in most regulated sectors.
We're already certified to ISO 9001 — how much extra work is ISO 42001?
Significantly less than a standalone ISO 42001 certification. ISO 42001 uses the same High-Level Structure (HLS) as ISO 9001, which means your context analysis, leadership structure, document control, internal audit framework, and management review process are already in place — or need only minor extension to cover AI-specific requirements. The incremental effort focuses on AI-specific additions: your AI inventory, AI risk register, data governance procedures, and Annex A controls. Clients with existing ISO 9001 certifications typically complete ISO 42001 certification in 3–5 months rather than 7–9.
Which certification body should we use?
DeepSynergy is certification-body agnostic — we work with any accredited certification body of your choice. In Southeast Asia and the Middle East we frequently work alongside PECB, SGS, Bureau Veritas, and TÜV Rheinland. In Europe, Lloyd's Register, BSI, and DNV are common choices. If you already have a relationship with a certification body from your ISO 9001 or ISO 14001 work, using the same body for ISO 42001 often reduces fees and simplifies the audit scheduling. We'll advise on the best fit for your region, sector, and existing certifications — with no financial relationship with any certification body.
What if we only use off-the-shelf AI tools — does ISO 42001 apply to us?
Yes — ISO 42001 explicitly covers organizations that use AI systems, not just those that develop them. If your staff use tools like Microsoft Copilot, ChatGPT, AI-powered recruitment screening, AI-assisted medical diagnosis, or any other AI system in their work — your organization is within scope. In fact, organizations using AI without building it often have the largest governance gaps, because they haven't thought systematically about what data they're feeding into vendor AI systems, what decisions those systems are influencing, or what oversight mechanisms are in place. These are exactly the risks ISO 42001 addresses.
How much does ISO 42001 certification cost?
Total cost depends on organization size, complexity of your AI portfolio, existing certifications, and which certification body you choose. Costs have two components: (1) DeepSynergy's consulting fees for gap analysis, system design, internal audit, and audit accompaniment — scoped based on your situation; and (2) the certification body's audit fees — typically $3,000–$12,000 depending on organization size. We provide a fixed-fee quote after an initial discovery call, so there are no surprises. For organizations with existing ISO certifications, total consulting fees are typically 30–40% lower due to shared infrastructure. Book a free 30-minute call and we'll give you a realistic estimate.

Ready to start your ISO 42001 journey?

Book a free 30-minute discovery call. We'll assess your current AI posture, estimate your timeline, and give you a realistic quote — no commitment required.