When do AI Act obligations apply? Deadlines after the 2026 amendment
Contents · 7 chapters
SHORT ANSWER
The AI Act, Regulation (EU) 2024/1689, entered into force on 1 August 2024 and applies in stages. The Article 5 prohibitions and the Article 4 AI literacy obligation have applied since 2 February 2025. The rules for general-purpose AI models and the penalty framework have applied since 2 August 2025. Most of the rest of the regulation, including the Article 50 transparency rules, applies from 2 August 2026. Two new prohibitions, on non-consensual intimate deepfakes and child sexual abuse material, apply from 2 December 2026. Regulation (EU) 2026/1744 postponed the high-risk obligations to 2 December 2027 for Annex III uses and to 2 August 2028 for products under Annex I.
The timeline after the 2026 amendment
| Date | What starts to apply | Provision |
|---|---|---|
| 1 Aug 2024 | The regulation enters into force. | Art. 113 |
| 2 Feb 2025 | General provisions, AI literacy (Article 4) and the Article 5 prohibitions. | Art. 113(a) |
| 2 Aug 2025 | Rules for general-purpose AI models (Chapter V), notifying authorities and notified bodies, governance including the AI Office, and penalties (Chapter XII) except Article 101. | Art. 113(b) |
| 27 Jul 2026 | Amending Regulation (EU) 2026/1744 enters into force. The changes to related product legislation in Articles 102 to 110 apply from this day. | Art. 4 of Regulation (EU) 2026/1744, Art. 113(d) |
| 2 Aug 2026 | General application of the rest of the regulation: transparency under Article 50, Commission fines for providers of general-purpose AI models under Article 101, and other provisions without a separate date. | Art. 113 |
| 2 Dec 2026 | Two new prohibitions: non-consensual intimate deepfakes and child sexual abuse material. End of the transition period for machine-readable marking by generators that were on the market before 2 Aug 2026. | Art. 113(a), Art. 111(4) |
| 2 Dec 2027 | Obligations for high-risk systems under Annex III, for example recruitment, credit scoring and education. Originally 2 Aug 2026. | Art. 113(c)(i) |
| 2 Aug 2028 | Obligations for high-risk systems in products under Annex I, for example medical devices or toys. Originally 2 Aug 2027. | Art. 113(c)(ii) |
| 2 Aug 2030 | Final deadline for high-risk systems intended for public authorities that were placed on the market earlier. | Art. 111(2) |
What the 2026/1744 amendment changed
The amendment, adopted on 8 July 2026 and published on 24 July 2026, entered into force on 27 July 2026. It changed the timeline in three ways:
- High-risk obligations postponed. Annex III from 2 August 2026 to 2 December 2027, Annex I from 2 August 2027 to 2 August 2028. According to the recitals of the amendment, the reason was the delay in harmonised standards.
- New prohibitions with their own date. Points (ba) and (bb) of Article 5 apply from 2 December 2026.
- A transition period for content marking. Generators placed on the market before 2 August 2026 must meet the machine-readable marking requirement by 2 December 2026.
A postponement is not a repeal. The high-risk obligations kept their full scope; only the date from which they apply moved.
What this means for an ordinary company
A company that uses AI but does not develop it has three things to do today: not use AI for prohibited practices, take measures supporting AI literacy, and since 2 August 2026 meet the transparency rules for deepfakes and emotion recognition. High-risk use, such as screening job applicants, has until 2 December 2027. How the Article 4 obligation translates into training by role is covered in AI training for companies: content, roles and the Article 4 obligation, and the full picture in the EU AI Act guide.
Sources and limits
The timeline is based on Articles 111 and 113 of Regulation (EU) 2024/1689 in the consolidated text of 27 July 2026, which includes the amendment by Regulation (EU) 2026/1744, verified on EUR-Lex on 22 September 2026. The consolidated text has no legal effect of its own; only the texts published in the Official Journal are binding. This answer replaces the version of 25 May 2026, which predated the amendment and gave wrong dates for general-purpose AI models, high-risk systems and biometrics.
Frequently asked questions
Does the AI Act already apply?
Yes. The regulation has been in force since 1 August 2024 and its parts apply in stages. As of 29 September 2026, the Article 5 prohibitions, the AI literacy obligation, the rules for general-purpose AI models, the penalty framework and, since 2 August 2026, the Article 50 transparency rules all apply. The high-risk obligations and the two new prohibitions added by the amendment do not apply yet.
Has the AI Act been postponed?
Only some deadlines moved. Regulation (EU) 2026/1744, in force since 27 July 2026, postponed the high-risk obligations for Annex III uses from 2 August 2026 to 2 December 2027 and for products under Annex I from 2 August 2027 to 2 August 2028. Providers of content generators already on the market received until 2 December 2026 for machine-readable marking. The prohibitions, AI literacy, the rules for general-purpose models and the other transparency obligations were not postponed.
From when must chatbots and AI-generated content be labelled?
The Article 50 transparency obligations apply from 2 August 2026. Providers of content generators that were on the market before that date have until 2 December 2026 to mark outputs in a machine-readable way under Article 50(2).
From when do the rules for ChatGPT and other large models apply?
The obligations of providers of general-purpose AI models under Chapter V have applied since 2 August 2025. Fines that the European Commission imposes on them directly under Article 101 can be imposed from 2 August 2026. Models placed on the market before 2 August 2025 have until 2 August 2027 under Article 111(3). These obligations fall on the model providers, not on companies that use the models.
What applies to high-risk systems placed on the market before the postponed date?
Under Article 111(2) as amended, the regulation applies to them only if their design changes significantly from the relevant date. High-risk systems intended for use by public authorities must comply in any case by 2 August 2030.
SOURCES AND VERIFICATION
- Regulation (EU) 2024/1689, Article 113 Entry into force and application (consolidated text of 27 July 2026, verified 22 September 2026)
- Regulation (EU) 2026/1744 amending Regulation (EU) 2024/1689
- Regulation (EU) 2024/1689, Article 111 AI systems already placed on the market or put into service (consolidated text)