EU AI Act

When it applies

Every date on which a EU AI Act obligation first bites, and what lands on each. Where a provision applies on more than one date, it appears on all of them.

§ 1 — How to read this

4 dates carry an obligation in this collection, 3 of them still ahead. Dates are taken from each obligation's own applicability column rather than restated here, so this page and the provision pages cannot disagree. Where a date is not self-explanatory — a transitional provision, a deferred application, a later amendment moving it — the reason is recorded against it.

§ 2 — Ahead

Still to come

2 December 2026
- Article 50 — Transparency obligations for providers and deployers of certain AI systems
  Article 50(2) only, for providers of generative systems placed on the market before 2 August 2026. A four-month transitional inserted by the Digital Omnibus. Systems placed on the market after that date owed the marking duty immediately.
2 December 2027
- Article 10 — Data and data governance
  Annex III standalone high-risk systems. Moved from 2 August 2026 by the Digital Omnibus — a sixteen-month extension driven by undesignated national authorities and the absence of harmonised standards, not by any relaxation of the Section 2 requirements themselves.
- Article 15 — Accuracy, robustness and cybersecurity
  Annex III standalone high-risk systems. Moved from 2 August 2026 by the Digital Omnibus — a sixteen-month extension driven by undesignated national authorities and the absence of harmonised standards, not by any relaxation of the Section 2 requirements themselves.
- Article 26 — Obligations of deployers of high-risk AI systems
  Annex III standalone high-risk systems. Moved from 2 August 2026 by the Digital Omnibus — a sixteen-month extension driven by undesignated national authorities and the absence of harmonised standards, not by any relaxation of the Section 2 requirements themselves.
- Article 27 — Fundamental rights impact assessment for high-risk AI systems
  Annex III standalone high-risk systems. Moved from 2 August 2026 by the Digital Omnibus — a sixteen-month extension driven by undesignated national authorities and the absence of harmonised standards, not by any relaxation of the Section 2 requirements themselves.
- Article 9 — Risk management system
  Annex III standalone high-risk systems. Moved from 2 August 2026 by the Digital Omnibus — a sixteen-month extension driven by undesignated national authorities and the absence of harmonised standards, not by any relaxation of the Section 2 requirements themselves.
2 August 2028
- Article 10 — Data and data governance
  Annex I embedded high-risk systems — medical devices, machinery, vehicles — where AI Act requirements fold into the existing sectoral conformity assessment. Moved from 2 August 2027.
- Article 15 — Accuracy, robustness and cybersecurity
  Annex I embedded high-risk systems — medical devices, machinery, vehicles — where AI Act requirements fold into the existing sectoral conformity assessment. Moved from 2 August 2027.
- Article 26 — Obligations of deployers of high-risk AI systems
  Annex I embedded high-risk systems — medical devices, machinery, vehicles — where AI Act requirements fold into the existing sectoral conformity assessment. Moved from 2 August 2027.
- Article 9 — Risk management system
  Annex I embedded high-risk systems — medical devices, machinery, vehicles — where AI Act requirements fold into the existing sectoral conformity assessment. Moved from 2 August 2027.

§ 3 — In force

Already in force

2 August 2026
- Article 50 — Transparency obligations for providers and deployers of certain AI systems
  The general application date. Chapter IV was not deferred by the Digital Omnibus, which moved only Chapter III Sections 1 to 3 — so these duties are in force while the high-risk regime is not.