Skip to main content
—
Framework library
Framework module · eu-ai-act

EU AI Act

The EU's risk-tiered regime for AI systems — prohibited practices, high-risk obligations, general-purpose model duties, and transparency requirements.

Regulation (EU) 2024/1689 · European Commission — AI Act · published 2024-07-12

Standing today
Directory entry

00Answer

from the registry record
What is EU AI Act?
The EU's risk-tiered regime for AI systems — prohibited practices, high-risk obligations, general-purpose model duties, and transparency requirements.
Who does EU AI Act apply to?
EU AI Act applies to AI, technology, regulated industries, EU, per European Commission — AI Act.
What is the current version of EU AI Act?
The current edition is Regulation (EU) 2024/1689, issued by European Commission — AI Act and published 2024-07-12. Source: https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai.
What does an assessment under EU AI Act require?
No control catalog has been ingested for EU AI Act yet — the registry tracks it as roadmap (Modeled framework: we model the regime — control families and at least one crosswalk map on disk — but no source-pinned control catalog is ingested), so an assessment under this framework currently requires the authority's own catalog rather than a ControlFrame-parsed one.

01Standing

Directory entry

Authority and version facts, with a parsed public catalog when available; no tenant blueprint or executable evidence method is activated.

EU AI Act is tracked in the registry — authority, version ledger, verification — and nothing else is modelled for it yet. Import its catalog to begin.

Freely published; obligations not modelled as a control catalog.

Request this frameworkAuthority source (opens in a new tab)

02Registry record

checked 2026-09-06
Registry status
Roadmap · modelledModeled framework: we model the regime — control families and at least one crosswalk map on disk — but no source-pinned control catalog is ingested.
Control units
Not on record — no catalog ingested and no authority-published count cited.
Control families
Prohibited practices · High-risk AI systems · General-purpose AI models · Transparency obligations · Risk management system
Applies to
AI · technology · regulated industries · EU
Verification
Primary — the issuing body's own page was read and states this version. checked 2026-09-06
Pending change
Regulation (EU) 2026/1744 (the 'Digital Omnibus on AI'), adopted 2026-07-08 and published in the OJ 2026-07-24, entered into force 2026-07-27, amending Regulation (EU) 2024/1689 (confirmed directly against the regulation's own EUR-Lex text). It defers Annex III high-risk obligations (Article 6(2) systems) to 2027-12-02 and Annex I embedded-product obligations (Article 6(1) systems) to 2028-08-02. Article 50 transparency duties began applying 2026-08-02, with a four-month transitional allowance for systems already placed on the market before that date. Treat the Omnibus as amending legislation, not a replacement framework.Expected: 2027-12-02 (Annex III high-risk)

03Version ledger

1 edition
Regulation (EU) 2024/1689Current edition2024-07-12

04Authority intelligence

reviewed 2026-08-30

Curated primary-source signals connected to this registry record. An authority change creates review work; it does not silently change tenant posture, evidence credit, or prior decisions.

  1. Regulatory milestone

    EU AI Act enforcement is now a staged operating calendar.

    Commission enforcement powers apply to provisions already in force, while transparency duties apply from August 2026 and the amended high-risk deadlines move to December 2027 and August 2028 by system class.

    Operating move

    Bind each AI system to its provider or deployer role, applicable article, system class, evidence owner, and effective date instead of using one undifferentiated readiness score.

    Read analysisEuropean Commission (opens in a new tab)

05Change history

Full tracker
  1. 2027-12-02
    EU AI Act: Regulation (EU) 2026/1744 (the 'Digital Omnibus on AI'), adopted 2026-07-08 and published in the OJ 2026-07-24, entered into force 2026-07-27, amending Regulation (EU) 2024/1689 (confirmed directly against the regulation's own EUR-Lex text). It defers Annex III high-risk obligations (Article 6(2) systems) to 2027-12-02 and Annex I embedded-product obligations (Article 6(1) systems) to 2028-08-02. Article 50 transparency duties began applying 2026-08-02, with a four-month transitional allowance for systems already placed on the market before that date. Treat the Omnibus as amending legislation, not a replacement framework.

06Related frameworks

scored from registry facts
  1. ISO/IEC 420012023

    Also applies to AI · The first certifiable management system standard for AI — the ISO 27001 shape applied to how an organization builds and operates AI systems.

  2. NIST AI Risk Management FrameworkAI RMF 1.0

    Also applies to AI · NIST's voluntary, cross-sector reference for governing AI risk, structured as Govern, Map, Measure, and Manage. NIST AI 600-1 is a companion profile for generative AI, not a replacement version or certification.

  3. ISO/IEC 238942023

    Also applies to AI · International guidance for integrating AI-specific risk management into organizations that develop, provide, deploy, or use AI systems. It complements ISO/IEC 42001 and is guidance, not a standalone certification.

  4. Texas Responsible AI Governance ActHB 149 (TRAIGA)

    Also applies to AI · Texas's AI law, narrowed before passage to intent-based prohibitions plus government-use rules, with a regulatory sandbox and AG enforcement.

See control-level overlap in the Control Atlas
EU AI Act | ControlFrame