EU-AI-Act
Regulation (EU) 2024/1689 — Artificial Intelligence Act
- I. General Provisions
- Art. 1. Subject matter ref
- Art. 2. Scope ref
- Art. 3. Definitions ref
- Art. 4. AI literacy ref
- II. Prohibited AI Practices
- Art. 5. Prohibited artificial intelligence practices ref
- III. High-Risk AI Systems
- Ch. 1 — Classification of AI Systems as High-Risk
- Art. 6. Classification rules for high-risk AI systems (7)
- Art. 7. Amendments to Annex III (12)
- Ch. 2 — Requirements for High-Risk AI Systems
- Art. 8. Compliance with the requirements (5)
- Art. 9. Risk management system (15)
- Art. 10. Data and data governance (20)
- Art. 11. Technical documentation (7)
- Art. 12. Record-keeping (8)
- Art. 13. Transparency and provision of information to deployers (14)
- Art. 14. Human oversight (11)
- Art. 15. Accuracy, robustness and cybersecurity (9)
- Ch. 3 — Obligations of Providers and Deployers of High-Risk AI Systems and Other Parties
- Art. 16. Obligations of providers of high-risk AI systems (12)
- Art. 17. Quality management system (16)
- Art. 18. Documentation keeping (6)
- Art. 19. Automatically generated logs (2)
- Art. 20. Corrective actions and duty of information (5)
- Art. 21. Cooperation with competent authorities (3)
- Art. 22. Duty of providers of high-risk AI systems to inform (2)
- Art. 23. Obligations of importers (12)
- Art. 24. Obligations of distributors (10)
- Art. 25. Responsibilities along the AI value chain (9)
- Ch. 4 — Obligations of Deployers of High-Risk AI Systems
- Art. 26. Obligations of deployers of high-risk AI systems (17)
- Art. 27. Fundamental rights impact assessment for high-risk AI systems (10)
- Ch. 5 — Notifying Authorities and Notified Bodies
- Art. 28. Notifying authorities (8)
- IV. Transparency Obligations for Providers and Deployers of Certain AI Systems
- Art. 50. Transparency obligations for providers and deployers of certain AI systems (9)
- V. General-Purpose AI Models
- Ch. 1 — Classification Rules
- Art. 51. Classification of general-purpose AI models as general-purpose AI models with systemic risk (4)
- Ch. 2 — Obligations for Providers of General-Purpose AI Models
- Art. 53. Obligations for providers of general-purpose AI models (6)
- Art. 54. Authorised representatives of providers of general-purpose AI models (11)
- Art. 55. Obligations for providers of general-purpose AI models with systemic risk (6)
- Art. 56. Codes of practice (8)
- VIII. Post-Market Monitoring, Information Sharing and Market Surveillance
- Ch. 1 — Post-Market Monitoring
- Art. 72. Post-market monitoring by providers and post-market monitoring plan for high-risk AI systems (7)
- Ch. 2 — Sharing of Information on Serious Incidents
- Art. 73. Reporting of serious incidents (12)
- X. Codes of Conduct and Guidelines
- Art. 95. Codes of conduct for voluntary application of specific requirements (6)
- XII. Penalties
- Art. 99. Penalties (8)
- Art. 100. Administrative fines on Union institutions, bodies, offices and agencies (7)
- Art. 101. Penalties for providers of general-purpose AI models (4)
- Annex I. Union Harmonisation Legislation Listed in Article 6(1)
- Annex III. High-Risk AI Systems Referred to in Article 6(2)
- Annex IV. Technical Documentation Referred to in Article 11(1)
Title I — General Provisions
Title II — Prohibited AI Practices
Title III — High-Risk AI Systems
Chapter 1 — Classification of AI Systems as High-Risk
Chapter 2 — Requirements for High-Risk AI Systems
Chapter 3 — Obligations of Providers and Deployers of High-Risk AI Systems and Other Parties
Chapter 4 — Obligations of Deployers of High-Risk AI Systems
Chapter 5 — Notifying Authorities and Notified Bodies
Title IV — Transparency Obligations for Providers and Deployers of Certain AI Systems
Article 50. Transparency obligations for providers and deployers of certain AI systems
2 obligations
EU-AIA-50-08
Transparency
Transparency information timing and manner
Information required under paragraphs 1 to 4 must be provided to natural persons in a clear and distinguishable manner a
EU-AIA-50-09
Transparency
AI-generated public interest text disclosure
Deployers of AI systems that generate or manipulate text published with the purpose of informing the public on matters o
Title V — General-Purpose AI Models
Chapter 1 — Classification Rules
Article 51. Classification of general-purpose AI models as general-purpose AI models with systemic risk
4 obligations
EU-AIA-51-01
Requirement
Classify general-purpose AI models based on high impact capabilities
General-purpose AI model providers must classify their models as having systemic risk if they have high impact capabilit
EU-AIA-51-02
Requirement
Classify models based on Commission decision regarding equivalent capabilities
General-purpose AI model providers must classify their models as having systemic risk when the Commission decides (ex of
EU-AIA-51-03
Requirement
Apply presumption of high impact capabilities for compute threshold models
General-purpose AI model providers must presume their model has high impact capabilities when the cumulative compute use
EU-AIA-51-04
Requirement
Commission obligation to adopt delegated acts for threshold amendments
The Commission shall adopt delegated acts according to Article 97 to amend the thresholds in paragraphs 1 and 2, and sup
Chapter 2 — Obligations for Providers of General-Purpose AI Models
Article 53. Obligations for providers of general-purpose AI models
6 obligations
EU-AIA-53-01
Documentation
Maintain technical documentation of general-purpose AI model
Draw up and keep up to date technical documentation of the model, including training and testing process and evaluation
EU-AIA-53-02
Transparency
Provide model information to downstream AI system providers
Draw up, keep up to date and make available information and documentation to providers of AI systems who intend to integ
EU-AIA-53-03
Requirement
Implement copyright compliance policy
Put in place a policy to comply with Union law on copyright and related rights, particularly to identify and comply with
EU-AIA-53-04
Transparency
Publish training data summary
Draw up and make publicly available a sufficiently detailed summary about the content used for training of the general-p
EU-AIA-53-05
Requirement
Cooperate with regulatory authorities
Cooperate as necessary with the Commission and national competent authorities in the exercise of their competences and p
EU-AIA-53-06
Conformity
Demonstrate compliance through approved means
When not adhering to approved codes of practice or European harmonised standards, demonstrate alternative adequate means
Article 54. Authorised representatives of providers of general-purpose AI models
11 obligations
EU-AIA-54-01
Requirement
Appoint authorised representative before market placement
Providers established in third countries must appoint, by written mandate, an authorised representative established in t
EU-AIA-54-02
Requirement
Enable authorised representative to perform mandated tasks
The provider must enable its authorised representative to perform all tasks specified in the written mandate received fr
EU-AIA-54-03
Requirement
Perform tasks specified in mandate
The authorised representative must perform all tasks specified in the mandate received from the provider.
EU-AIA-54-04
Transparency
Provide mandate copy to AI Office upon request
The authorised representative must provide a copy of the mandate to the AI Office upon request, in one of the official l
EU-AIA-54-05
Requirement
Verify technical documentation and obligation fulfillment
The mandate must empower the authorised representative to verify that the technical documentation specified in Annex XI
EU-AIA-54-06
Documentation
Keep technical documentation and provider contact details for 10 years
The mandate must empower the authorised representative to keep a copy of the technical documentation specified in Annex
EU-AIA-54-07
Transparency
Provide compliance information and documentation to AI Office
The mandate must empower the authorised representative to provide the AI Office, upon a reasoned request, with all the i
EU-AIA-54-08
Requirement
Cooperate with AI Office and competent authorities
The mandate must empower the authorised representative to cooperate with the AI Office and competent authorities, upon a
EU-AIA-54-09
Requirement
Empower representative to be addressed by authorities
The mandate must empower the authorised representative to be addressed, in addition to or instead of the provider, by th
EU-AIA-54-10
Requirement
Terminate mandate if provider acts contrary to obligations
The authorised representative must terminate the mandate if it considers or has reason to consider the provider to be ac
EU-AIA-54-11
Reporting
Inform AI Office of mandate termination and reasons
When terminating the mandate due to provider non-compliance, the authorised representative must immediately inform the A
Article 55. Obligations for providers of general-purpose AI models with systemic risk
2 obligations
EU-AIA-55-01
Requirement
Perform model evaluation with standardized protocols and adversarial testing
Conduct model evaluation using standardized protocols and tools reflecting state of the art, including adversarial testi
EU-AIA-55-02
Risk Management
Assess and mitigate systemic risks from model development and deployment
Assess and mitigate possible systemic risks including their sources that may arise from development, placing on market,