AI governance under the EU AI Act and ISO/IEC 42001 – secure, demonstrable, integrated
Is your company using or developing artificial intelligence? We create the framework for it: from the AI inventory and risk classification under the EU AI Act to a certifiable AI management system – embedded in your existing ISMS and privacy management.
- EU AI Act
- ISO/IEC 42001
- EN 18286
- ISO/IEC 42005
- ISO/IEC 23894
- German KI-MIG
What is AI governance?
AI governance is the framework a company uses to steer the use and development of AI: Who may introduce which AI systems? How are risks to people, data and business processes assessed? Who monitors the systems in operation, and how is this documented?
Without such a framework, “shadow AI” emerges: business units use chatbots, copilots and AI features in standard software without management, information security or data protection being aware of it. This is not only a security and privacy risk – since the EU AI Act, it is also a liability issue for management.
Good AI governance connects three levels: law (EU AI Act, GDPR, copyright and employment law), standards (above all ISO/IEC 42001) and processes that are actually lived in the company. Connecting these is our strength.
EU AI Act timeline
Regulation (EU) 2024/1689 applies in stages (in detail on our page on the EU AI Act). Following the Digital Omnibus, all application dates are fixed:
- 2 Feb 2025 Prohibited AI practices (Art. 5) and AI literacy (Art. 4)
- 2 Aug 2025 Obligations for providers of general-purpose AI models (GPAI)
- 27 Jul 2026 Digital Omnibus (Regulation (EU) 2026/1744) in force – Art. 4 revised: companies must take measures to promote AI literacy
- 29 Jul 2026 German AI Market Surveillance and Innovation Act (KI-MIG): the Federal Network Agency becomes market surveillance authority, single point of contact and complaints body
- 2 Aug 2026 General application, transparency obligations (Art. 50), penalties and supervision
- October 2026 EN 18286 published in the Official Journal of the EU as the first harmonised standard
- 2 Dec 2027 Obligations for high-risk AI under Annex III (e.g. recruitment, creditworthiness, critical infrastructure) including deployer obligations and fundamental rights impact assessment
- 2 Aug 2028 Obligations for high-risk AI in regulated products under Annex I (e.g. machinery, medical devices)
Fines under Art. 99: up to EUR 35 million or 7 % of worldwide annual turnover for prohibited practices, up to EUR 15 million or 3 % for breaches of other obligations. The postponement does not change the fact that inventory, classification and processes must be set up now – high-risk requirements cannot be implemented in a few weeks.
Standards for AI governance at a glance
The EU AI Act states what must be achieved. The standards describe how to implement it demonstrably. These standards form the foundation of our consulting:
ISO/IEC 42001:2023
The world's first certifiable standard for an AI management system (AIMS). Same harmonised structure as ISO/IEC 27001 – ideal for integration into your ISMS. Annex A contains the AI-specific controls.
EN 18286:2026
First harmonised standard for the EU AI Act: quality management system under Art. 17 for providers of high-risk AI. Gives a presumption of conformity for the requirements it covers and builds on ISO/IEC 42001 or ISO 9001.
ISO/IEC 42005:2025
Guidance on AI system impact assessment across the entire lifecycle – effects on individuals, groups and society. Operational basis for the assessments under ISO/IEC 42001.
ISO/IEC 23894:2023
AI-specific risk management based on ISO 31000. Methodological basis for the risk management system under Art. 9 of the EU AI Act.
ISO/IEC 42006:2025
Requirements for certification bodies auditing AI management systems against ISO/IEC 42001. Ensures that your certificate comes from competent auditors.
ISO/IEC 38507 · 22989 · 5338
Governance responsibilities of the governing body for the use of AI (38507), common AI terminology (22989) and processes for the AI lifecycle (5338).
ISO/IEC 27001 · 27701
AI systems are information assets: the information security and privacy information management systems form the basis on which the AIMS builds – without duplicate structures.
NIST AI RMF · BSI AIC4
NIST AI Risk Management Framework (including the generative AI profile) for US business and group requirements; the German BSI AIC4 criteria catalogue for AI cloud services in supplier assessment.
Further harmonised standards on risk management, data quality, logging, transparency, human oversight, accuracy, robustness and cybersecurity are being developed by CEN/CENELEC JTC 21. Our regulatory monitoring keeps you up to date.
Our offer: CONSUVATION AI Governance
We do not deliver empty templates but a ready-made, field-proven process and management model that we tailor to your company together with you. Four modules – individually or as a complete package:
AI Check
Inventory in just a few days.
- AI inventory including shadow AI and AI in standard software
- Role clarification: provider, deployer, importer
- Risk classification under the EU AI Act
- Gap analysis against ISO/IEC 42001
Result: AI register, risk map, prioritised roadmap
Governance framework
Clear rules for the use of AI.
- AI policy and usage rules (e.g. for generative AI)
- Roles: AI officer, AI board, business owners
- Approval process for new AI use cases
- AI literacy measures under Art. 4
Result: adopted policy, approval workflow, training plan
AI management system
AIMS under ISO/IEC 42001 – integrated.
- Processes, policies and templates from our management model
- Risk management under ISO/IEC 23894
- Impact assessment under ISO/IEC 42005, linked to the DPIA
- Managed with our GRC tool (ISMS, privacy and AI)
Result: certifiable AIMS, support through to certification
Operation & evidence
Stay compliant over time.
- External AI officer – or support for your internal one
- Internal audits under ISO/IEC 42001
- Regulatory monitoring: EU AI Act, KI-MIG, guidelines
- Target-group-specific e-learning on AI literacy
Result: up-to-date legal register, audit reports, KPI dashboard
Conformity by December 2027
Do you develop AI systems for HR, financial services, critical infrastructure or as a safety component of products? We prepare you for the conformity assessment:
- Quality management system under EN 18286 based on ISO/IEC 42001 or ISO 9001
- Risk management system (Art. 9) and data governance (Art. 10)
- Technical documentation under Annex IV
- Human oversight, logging, post-market monitoring
Using AI safely
Most companies use AI but do not develop it. For you, pragmatic solutions count:
- Assessment of AI suppliers and AI features in cloud services
- Transparency obligations under Art. 50 (e.g. chatbots, AI-generated content)
- Fundamental rights impact assessment (Art. 27) where required
- Alignment with data protection, works council and information security
Ready to use: our AI management system and our AI risk assessments
You do not start from scratch. CONSUVATION has developed a complete AI management system under ISO/IEC 42001 and its own AI risk assessments based on international standards. We adapt both to your company instead of reinventing them – saving time, effort and rounds of coordination.
AI management system under ISO/IEC 42001
A ready-made management model with all required elements:
- AI policy, AI guidelines and role model
- Processes for the entire AI lifecycle
- ISO/IEC 42001 document library with 20 documents in German and English
- Templates for AI register, approvals and evidence
- Implementation of all Annex A controls
- Integrable into ISMS (ISO/IEC 27001) and PIMS (ISO/IEC 27701)
- Managed in the AI module of our GRC platform (AI Act Manager)
AI risks in the ISMS
Our risk assessment extends your ISMS with AI-specific threats, based on ISO/IEC 27001, ISO/IEC 23894 and ISO 31000:
- Data leakage via AI services and prompts
- Prompt injection, data poisoning, model manipulation
- Incorrect or hallucinated results in processes
- Dependence on AI suppliers and cloud models
- Shadow AI in business units
Result: AI risks directly in the existing ISMS risk register
Risk assessment under the EU AI Act
Structured analysis along the obligations of the EU AI Act:
- Check for prohibited practices (Art. 5)
- Risk classification: high-risk under Annex I and III, transparency, minimal risk
- Risk management system for high-risk AI (Art. 9)
- Fundamental rights impact assessment (Art. 27)
- Impact assessment under ISO/IEC 42005, linked to the DPIA under Art. 35 GDPR
Result: documented classification per AI system with action plan · Free risk class self-check
One system instead of three: AI governance in an integrated management system
ISO/IEC 42001, ISO/IEC 27001 and ISO/IEC 27701 share the same harmonised structure. We consistently use these commonalities – your existing processes are extended, not duplicated:
| EU AI Act requirement | Implementation in standards | CONSUVATION module |
|---|---|---|
| Art. 4 AI literacy | ISO/IEC 42001 cl. 7.2/7.3 | Training plan, e-learning modules |
| Art. 5 Prohibitions | Approval process, Annex A.5 | AI Check, use case approval |
| Art. 9 Risk management | ISO/IEC 23894, ISO 31000 | Risk methodology in the GRC tool |
| Art. 10 Data governance | ISO/IEC 42001 Annex A.7 | Data and model policy |
| Art. 11 / Annex IV Documentation | ISO/IEC 42001 cl. 7.5 | Documentation templates, AI register |
| Art. 17 Quality management | EN 18286, ISO/IEC 42001 | Management model for high-risk AI |
| Art. 27 Fundamental rights impact assessment | ISO/IEC 42005 | Impact assessment incl. DPIA |
| Art. 50 Transparency | ISO/IEC 42001 Annex A.8 | Labelling and information rules |
| Art. 72 Post-market monitoring | EN 18286, ISO/IEC 42001 cl. 9 | Monitoring, KPI dashboard, internal audits |
Do you already have an ISMS under ISO/IEC 27001 or are you implementing NIS2? Then a large part of the work is already done – the AI management system builds directly on it.
Why CONSUVATION?
- More than 25 years of experience with management systems, audits and certifications – from the early days of BS 7799 to today.
- First-hand standardisation know-how through participation in the ISO committees for ISO/IEC 27001 and ISO/IEC 27701.
- Certified expertise: CISA, CISM, CRISC and CGEIT – IT governance is our core business.
- Our own AI management system under ISO/IEC 42001 and our own standards-based AI risk assessments, ready to use.
- Solutions instead of empty templates: ready-made process and management models, managed with our GRC tool.
- Integrated thinking: AI, information security and data protection from a single source – without parallel structures.
- Very high success rate in audits and certifications of our clients from automotive, banking, insurance, pharma, industry, energy, IT and defence.
Frequently asked questions on AI governance
What is AI governance?
AI governance is the framework a company uses to steer the use and development of artificial intelligence: responsibilities, policies, approval processes, risk assessment, documentation and monitoring. It ensures that AI is used lawfully, securely and transparently.
Does the EU AI Act also apply to companies that only use AI?
Yes. Deployers of AI systems also have obligations, such as measures to promote AI literacy (Art. 4), compliance with the prohibitions (Art. 5) and transparency obligations (Art. 50). For high-risk AI, extensive deployer obligations apply from 2 December 2027.
Is ISO/IEC 42001 certification mandatory?
No. ISO/IEC 42001 is voluntary. A certified AI management system is, however, the recognised proof of responsible AI use towards customers, authorities and partners, and a solid basis for the obligations of the EU AI Act.
What role does EN 18286 play?
EN 18286:2026 is the first harmonised standard for the EU AI Act. It specifies the quality management system under Art. 17 for providers of high-risk AI and gives a presumption of conformity for the requirements it covers. It builds on ISO/IEC 42001 or ISO 9001.
When do the obligations for high-risk AI apply?
Following the amendment by Regulation (EU) 2026/1744 (Digital Omnibus), obligations for high-risk AI under Annex III apply from 2 December 2027, and for AI in regulated products under Annex I from 2 August 2028.
Can AI governance be integrated into an existing ISMS?
Yes. ISO/IEC 42001 follows the same harmonised structure as ISO/IEC 27001 and ISO/IEC 27701. Roles, document control, risk management, internal audits and management review are shared, so no parallel system is created.
How are AI risks assessed?
We use two risk assessments we developed on the basis of international standards: an AI risk assessment for the ISMS based on ISO/IEC 27001, ISO/IEC 23894 and ISO 31000, and a risk assessment under the EU AI Act that checks for prohibited practices, classifies AI systems by risk class and includes the impact assessment under ISO/IEC 42005 and Art. 27.
Start with the AI Check
Within a few days you will know which AI systems are in use, which obligations apply to you and which steps make sense next. Get in touch – personal, competent, straightforward.
Note: This page provides an overview of the legal situation and standards (as of November 2026) and does not constitute legal advice.