Article 9 of 1138%
Article 9: Risk management system
EU Official:
UpcomingAnnex III from 2 Dec 2027; Annex I from 2 Aug 2028
Title III: High-Risk AI Systems
Application dates
- : Requirements for high-risk systems listed in Annex III
- : Requirements for high-risk systems linked to Annex I
Article 9 is the risk-management foundation for high-risk AI. Organisations must not only identify risks, but also mitigate, test, monitor and document them.
Official text
||
Source: EUR-Lex, Regulation (EU) 2024/1689. Text reproduced verbatim.
Download AI Act (PDF)→Official guidance on this article
4- Commission Implementing Decision C(2025) 3871 of 23.6.2025 on a standardisation request to CEN and CENELEC as regards high-risk AI systems in support of Regulation (EU) 2024/1689 and repealing Implementing Decision C(2023) 3215Explicitly requests a standard for the risk management system so that providers can implement the open standard of Article 9 with a concrete methodologyPublishedImplementing actEuropean Commission23 Jun 2025
- ISO/IEC 42005:2025 Information technology - Artificial intelligence (AI) - AI system impact assessmentSupplements the risk management system of Article 9 with an explicit focus on impacts on individuals and groups, making the risk process broader than product safety alone.PublishedStandardISO/IEC JTC 1/SC 42, Dutch adoption by NEN1 Jun 2025
- EN ISO/IEC 23894:2024 Information technology - Artificial intelligence - Guidance on risk managementCurrently provides a usable methodology for the risk management system of Article 9, as a bridge while the European prescriptive standard prEN 18228 remains in draft phase.PublishedStandardISO/IEC JTC 1/SC 42, European adoption through CEN-CENELEC, Dutch adoption by NEN1 Feb 2024
- prEN 18228 Artificial intelligence - Risk management (draft European standard, public enquiry)Provides the process that providers must follow to meet the obligation under Article 9 to maintain a risk management system throughout the entire lifecycle, including risks to fundamental rights.DraftStandardCEN-CENELEC JTC 21, national consultation through NEN18 May 2026
What does this mean for you?
Provider+
You must establish a continuous risk management system that identifies, analyses and mitigates risks throughout the entire lifecycle. This is not a one-time assessment but an ongoing process. Document everything in your technical documentation (Art. 11).
Deployer+
Ensure you understand and monitor your provider's risk management system. Report deviations from expected behaviour to the provider. Maintain a log of incidents and deviations.
SME / Startup+
Start small: create a risk inventory of your AI use. What data goes in? What decisions come out? What could go wrong? Use our FRIA Generator as a starting point.
Public Sector+
Combine your AI risk management with existing frameworks such as the IAMA (Human Rights and Algorithms Impact Assessment). The AI Act requirements align with what many governments are already doing.
Compliance checklist
- ☐Risk management system established and documented
- ☐Risk identification and analysis performed
- ☐Risk mitigation measures implemented
- ☐Residual risks assessed and deemed acceptable
- ☐Testing procedures performed before deployment
- ☐Continuous monitoring system established
Want to save your progress? Create an account
Related recitals
Recital 65Primary
The risk-management system should consist of a continuous, iterative process that is planned and run throughout the entire lifecycle of a high-risk AI system. That process should be aimed at identifyi…
Recital 66Primary
Requirements should apply to high-risk AI systems as regards risk management, the quality and relevance of data sets used, technical documentation and record-keeping, transparency and the provision of…
Recital 67Primary
High-quality data and access to high-quality data plays a vital role in providing structure and in ensuring the performance of many AI systems, especially when techniques involving the training of mod…
Related tools
Related enforcement
No enforcement actions for this article yet. Follow developments via the Enforcement Tracker.
Related blog posts
Cross-references
Annexes
Frequently asked questions
What does the risk management system for AI entail?+
Article 9 requires a continuous, iterative risk management system that identifies, analyses, evaluates and mitigates risks throughout the entire lifecycle of the AI system.
Is a risk assessment one-time or ongoing?+
Risk management must be ongoing. It is not a one-time check but a continuous process that is regularly updated, especially when significant changes are made to the AI system.
Do SMEs also need to comply with Article 9 of the AI Act?+
Article 9 of the AI Act does not provide a general exemption for SMEs. However, the AI Act includes supportive measures and potentially lighter obligations for small and medium-sized enterprises, depending on their role in the AI value chain.
How does Article 9 of the AI Act relate to the GDPR?+
Article 9 of the AI Act complements the GDPR. While the GDPR protects personal data, the AI Act focuses on the safety and trustworthiness of AI systems. Organisations must comply with both regulations when their AI system processes personal data.
What are the deadlines for Article 9 of the AI Act?+
The AI Act follows a phased implementation. Prohibited AI practices apply from February 2025, obligations for high-risk AI systems from August 2026, and other provisions take effect gradually. The specific deadline for Article 9 depends on the category of the obligation.
Does Article 9 of the AI Act also apply to AI systems I purchase?+
Yes, Article 9 of the AI Act may also be relevant when you purchase AI systems. As a deployer, you have your own obligations under the AI Act, regardless of whether you developed the system yourself or purchased it from a provider.
What is the difference between provider and deployer under Article 9 of the AI Act?+
Under Article 9 of the AI Act, the provider is the entity that develops or places the AI system on the market, while the deployer is the entity that uses the system under its own authority. Both roles carry different obligations.
What documentation does Article 9 of the AI Act require?+
Article 9 of the AI Act requires that relevant documentation is maintained as part of the compliance process. This may include technical documentation, instructions for use, logs or declarations of conformity, depending on the classification of the AI system.
How often must the risk management system be updated?+
The risk management system must be a continuous, iterative process throughout the entire lifecycle of the AI system. It must be updated when substantial modifications are made, when new risk insights emerge, or when post-market monitoring yields new information.
What is the difference between a DPIA (GDPR) and the risk management system of Article 9?+
A DPIA under the GDPR focuses on risks to privacy and data protection. The Article 9 risk management system is broader and covers risks to health, safety and all fundamental rights. Both assessments are complementary and may partially overlap, but do not replace each other.
Do I need to explicitly document residual risks?+
Yes, Article 9 requires that residual risks remaining after risk mitigation measures are explicitly identified and documented. These residual risks must be assessed as acceptable and communicated to the deployer through the instructions for use.
Can I combine the AI Act risk management system with ISO 31000 or ISO 23894?+
Yes, existing risk management frameworks like ISO 31000 (risk management) and ISO/IEC 23894 (AI risk management) can serve as a foundation. However, you must ensure all specific Article 9 requirements are covered, including the focus on fundamental rights and the obligation to test with representative data.