EU AI Act application dates
The EU AI Act does not apply on a single date. Each obligation group below shows when it applies, read from the stored text of the provision it comes from, and each row names that provision. Where Regulation (EU) 2026/1744 moved a date, the row says so and gives the old one.
Last modified 2 August 2026: the most recent date on which a stored text behind this table was recorded or an obligation group began to apply. This date does not move back.
| Obligation group | Date | What it covers | Where the date comes from |
|---|---|---|---|
| first banned practices and general rules The Act's opening rules: general provisions, AI literacy duties, and the original banned-practices list. | Has applied since 2 February 2025 | Chapters I and II: general provisions, AI literacy (Article 4) and the original Article 5 prohibitions | Article 113, third paragraph, point (a), as replaced by Regulation (EU) 2026/1744, Article 1, point (40)(a). The date is unchanged; the operative text is now the amending one. Regulation (EU) 2026/1744, Article 1. |
| general-purpose AI model duties, governance and penalties Duties for general-purpose AI model providers, plus the Act's authorities, governance and penalties machinery. | Has applied since 2 August 2025 | Notifying authorities and notified bodies (Chapter III, Section 4), general-purpose AI model obligations (Chapter V), governance (Chapter VII), penalties (Chapter XII) and Article 78, excluding Article 101 | Article 113, third paragraph, point (b) of the base Regulation. Point (40) does not touch point (b). Article 113 of Regulation (EU) 2024/1689. |
| changes to other EU laws The changes the Act makes to other EU laws. | Has applied since 27 July 2026 | Articles 102 to 110: amendments to the sectoral Union legislation | Article 113, third paragraph, point (d), added by Regulation (EU) 2026/1744, Article 1, point (40)(c). A new phase that did not exist in the base act. Regulation (EU) 2026/1744, Article 1. |
| transparency duties Telling people they are interacting with AI, and marking AI-generated content. | Has applied since 2 August 2026 | Transparency obligations (Chapter IV, Article 50) | Article 113, second paragraph of the base Regulation: the residual date. Chapter IV is listed nowhere in the third paragraph, so the general application date catches it. Point (40) amends only the third paragraph and so does not move this date. Separately, point (39)(b) adds a transitional for systems already on the market. See the Article 50(2) transitional row; it does not change the date on this row. Article 113 of Regulation (EU) 2024/1689. |
| fines for general-purpose AI model providers Fines for general-purpose AI model providers. | Has applied since 2 August 2026 | Article 101: fines for providers of general-purpose AI models | Article 113, third paragraph, point (b) of the base Regulation applies Chapter XII from 2 August 2025 'with the exception of Article 101', which returns Article 101 to the residual date in the second paragraph. Same date as Article 50, different derivation, and distinct from the general-purpose AI obligations above, which have bound providers since 2025. Article 113 of Regulation (EU) 2024/1689. |
| conformity assessment, CE marking and registration of high-risk systems How a high-risk system is assessed, declared, CE-marked and registered. | Has applied since 2 August 2026 | Chapter III, Section 5: conformity assessment, the EU declaration of conformity, CE marking and registration | Article 113, second paragraph of the base Regulation: the residual date. The third paragraph names Chapter III Section 4 in point (b), and Chapter III Sections 1, 2 and 3 in point (c) as replaced by Regulation (EU) 2026/1744, Article 1, point (40)(b). It names Section 5 nowhere, so the general application date catches it. Article 113 of Regulation (EU) 2024/1689. |
| older AI components of large-scale EU IT systems Older AI components inside the EU's listed large-scale IT systems, already in service before the cut-off; they must comply by this date. | Applies from 31 December 2030 | Article 111(1): AI systems which are components of the large-scale IT systems established by the legal acts listed in Annex X and already placed on the market or put into service before the paragraph's cut-off | Article 111(1) of the base Regulation, which sets its own compliance deadline rather than deferring to Article 113. This product does not hold Annex X, so it cannot say which systems the cohort contains. Article 111 of Regulation (EU) 2024/1689. |
| older high-risk systems used by public authorities Older high-risk systems used by public authorities, already on the market before the high-risk rules applied; they must comply by this date. | Applies from 2 August 2030 | Article 111(2): providers and deployers of high-risk AI systems intended to be used by public authorities, where the system was placed on the market or put into service before Chapter III applies | Article 111(2) as replaced by Regulation (EU) 2026/1744, Article 1, point (39)(a). It composes with the Chapter III deferral: significant design changes from the date of application of Chapter III bring a legacy system in, and this date is the backstop that applies in any case. Regulation (EU) 2026/1744, Article 1. |
| older general-purpose AI models General-purpose AI models already on the market before their duties applied in August 2025; they must comply by this date. | Applies from 2 August 2027 | Article 111(3): providers of general-purpose AI models that were already on the market before the general-purpose AI obligations applied | Article 111(3) of the base Regulation, which sets its own compliance deadline for the legacy general-purpose model cohort. Article 111 of Regulation (EU) 2024/1689. |
| registration in the EU database The EU database where Annex III high-risk systems are registered. | Has applied since 2 August 2026 | Chapter VIII: the EU database for high-risk AI systems listed in Annex III | Article 113, second paragraph of the base Regulation: the residual date. No point of the third paragraph names Chapter VIII, and the closed Chapter and Section address introduced by Regulation (EU) 2026/1744, Article 1, point (40)(b) does not reach it. Article 113 of Regulation (EU) 2024/1689. |
| post-market monitoring, incident reporting and the right to an explanation Watching a system after launch, reporting serious incidents, and explaining individual decisions. | Has applied since 2 August 2026 | Chapter IX: post-market monitoring by providers, reporting of serious incidents, and the right to an explanation of individual decision-making | Article 113, second paragraph of the base Regulation: the residual date. No point of the third paragraph names Chapter IX. The superseded point (c) read 'Article 6(1) and the corresponding obligations in this Regulation', which was broad enough to carry these duties for an Annex I system; Regulation (EU) 2026/1744, Article 1, point (40)(b) replaced that formula with a closed Chapter and Section address, so they fall back to the residual date. Article 113 of Regulation (EU) 2024/1689. |
| newly added banned practices Banned practices added by the 2026 amendment to the Act. | Applies from 2 December 2026 | New Article 5 prohibitions: Article 5(1), first subparagraph, points (ba) and (bb), and Article 5(1a) and (1b) | Article 113, third paragraph, point (a), as replaced by Regulation (EU) 2026/1744, Article 1, point (40)(a). New prohibitions with their own later date, carved out of the Chapters I and II phase. Regulation (EU) 2026/1744, Article 1. |
| marking of AI-made content by systems already on the market Systems that generate synthetic content and were already on the market before Article 50 applied get until this date to mark their output as AI-made. | Applies from 2 December 2026 | Transitional: providers of AI systems generating synthetic audio, image, video or text content that were already placed on the market before the Act's general application date, compliance with Article 50(2) only | Article 111(4), added by Regulation (EU) 2026/1744, Article 1, point (39)(b). This is relief for a legacy cohort, not a postponement of Article 50: it reaches providers only (not deployers), systems generating synthetic content only, systems placed on the market before the general application date only, and Article 50(2) only. Article 50(1), (3) and (4) are untouched, and systems placed on the market from the general application date onward comply on that date. Regulation (EU) 2026/1744, Article 1. |
| high-risk duties for systems in the Annex III use areas The main high-risk duties, for systems high-risk because of where they are used: hiring, credit, education, essential services and the other Annex III areas. | Applies from 2 December 2027 · moved from 2 August 2026 by Regulation (EU) 2026/1744 | High-risk obligations, Chapter III Sections 1, 2 and 3 (except Article 6(5)): systems classified high-risk under Article 6(2) and Annex III | Article 113, third paragraph, point (c)(i), as replaced by Regulation (EU) 2026/1744, Article 1, point (40)(b). Previously caught by the residual date in the second paragraph. Regulation (EU) 2026/1744, Article 1. |
| high-risk duties for AI inside Annex I regulated products The main high-risk duties, for AI that is a safety component of a product already regulated under Annex I, such as machinery or medical devices. | Applies from 2 August 2028 · moved from 2 August 2027 by Regulation (EU) 2026/1744 | High-risk obligations, Chapter III Sections 1, 2 and 3 (except Article 6(5)): systems classified high-risk under Article 6(1) and Annex I | Article 113, third paragraph, point (c)(ii), as replaced by Regulation (EU) 2026/1744, Article 1, point (40)(b). The superseded point (c) applied Article 6(1) and its corresponding obligations from an earlier date. Regulation (EU) 2026/1744, Article 1. |
Sources
- Regulation (EU) 2024/1689 (Artificial Intelligence Act); Article 113; ELI http://data.europa.eu/eli/reg/2024/1689/oj; CELEX 32024R1689. Source: Regulation (EU) 2024/1689, © European Union, https://eur-lex.europa.eu, 1998–2024. Only EU legislation published in the electronic edition of the Official Journal of the European Union is deemed authentic.
- Regulation (EU) 2026/1744 (Digital Omnibus on AI); Regulation (EU) 2026/1744, Article 1; ELI http://data.europa.eu/eli/reg/2026/1744/oj; CELEX 32026R1744. Source: Regulation (EU) 2026/1744, © European Union, https://eur-lex.europa.eu, 1998–2026. Only EU legislation published in the electronic edition of the Official Journal of the European Union is deemed authentic. Adopted 8 July 2026; published OJ L, 2026/1744, 24.7.2026; IN FORCE since 27 July 2026 (Article 4, third day following publication). It AMENDS Regulation (EU) 2024/1689; read alongside the base act, which is pinned separately and unchanged. Reviewed by our founder on 15 September 2026: as at that date, he had personally reviewed these provisions against the official text; this is not external legal review.
- Regulation (EU) 2024/1689 (Artificial Intelligence Act); Article 111; ELI http://data.europa.eu/eli/reg/2024/1689/oj; CELEX 32024R1689. Source: Regulation (EU) 2024/1689, © European Union, https://eur-lex.europa.eu, 1998–2024. Only EU legislation published in the electronic edition of the Official Journal of the European Union is deemed authentic.
An automated reading of the official texts, not legal advice.
Perissos's compilation and annotations in this table and in its JSON are licensed under CC BY 4.0. The licence does not extend to the EU legal texts, which keep their EUR-Lex attribution. To attribute, use: Perissos, EU AI Act application dates, https://perissos.ai/eu-ai-act-dates