Since 2 August 2026 the AI Act applies in general: the Article 50 transparency obligations (chatbots, generated content, deepfakes) are in force, on top of the prohibitions and the AI literacy duty that have applied since 2 February 2025. The high-risk obligations, however, were postponed by Regulation (EU) 2026/1744, the "Digital Omnibus": to 2 December 2027 for Annex III systems and to 2 August 2028 for AI embedded in products covered by Annex I.
What did the Digital Omnibus change?
Regulation (EU) 2026/1744 was signed on 8 July 2026, published in the Official Journal on 24 July 2026 and entered into force on 27 July 2026. It amends Regulation (EU) 2024/1689 in several places. Recital 2 explains the delay by the late preparation of harmonised standards and the slow set-up of national authorities and conformity assessment bodies.
The main changes:
- High-risk postponed: Chapter III, Sections 1 to 3 (classification, requirements, obligations of providers and deployers) apply from 2 December 2027 for Annex III and 2 August 2028 for Annex I (Article 113).
- Grace period for marking: providers of generative systems placed on the market before 2 August 2026 have until 2 December 2026 to comply with Article 50(2) (Article 111(4)).
- New prohibitions: from 2 December 2026, Article 5 bans AI systems that generate intimate or sexually explicit images of an identifiable person without consent, and child sexual abuse material (points (ba) and (bb)).
- AI literacy reworded: Article 4 now requires providers and deployers to "take measures" to support their staff's AI literacy, without having to guarantee any specific level for each individual.
- Relief extended: several SME measures (simplified documentation, proportionate quality management, capped fines) now also cover small mid-caps (SMCs).
The postponement only covers Chapter III. It does not affect Article 50, the prohibitions or the obligations for general-purpose AI models.
What are the key dates?
| Date | What applies | Reference |
|---|---|---|
| 1 August 2024 | AI Act enters into force | Article 113 |
| 2 February 2025 | Definitions, AI literacy, prohibited practices | Chapters I and II, Articles 4 and 5 |
| 2 August 2025 | General-purpose AI models, governance, notified bodies, penalties regime (except Article 101) | Article 113(b) |
| 27 July 2026 | Omnibus enters into force | Regulation 2026/1744 |
| 2 August 2026 | General application: transparency (Article 50), market surveillance, Commission fining powers over general-purpose AI providers (Article 101) | Article 113 |
| 2 December 2026 | New prohibitions (Article 5(1)(ba) and (bb)); marking deadline for generative systems already on the market | Articles 113 and 111(4) |
| 2 August 2027 | Compliance deadline for general-purpose AI models placed on the market before 2 August 2025 | Article 111(3) |
| 2 December 2027 | Annex III high-risk systems (employment, credit, education, essential services…) | Article 113(c)(i) |
| 2 August 2028 | Annex I high-risk systems (safety components of regulated products) | Article 113(c)(ii) |
| 2 August 2030 | High-risk systems already in service and used by public authorities | Article 111(2) |
Other high-risk systems already on the market before their application date are only caught if their design changes significantly after that date (Article 111(2)).
What do I have to do right now?
The high-risk delay does not mean nothing applies today. Already:
- any chatbot or agent that interacts with people must tell them it is an AI (Article 50(1)); see is my chatbot covered?;
- synthetic content must be marked (Article 50(2)) and deepfakes disclosed by the deployer (Article 50(4));
- no Article 5 practice may be used (manipulation, exploiting vulnerabilities, social scoring and the rest of the list);
- AI literacy measures must be in place for staff (Article 4).
For systems that will fall under Annex III, the time until 2 December 2027 is for building the technical documentation, quality management system, logging and human oversight. The first published harmonised standard, EN 18286:2026 on quality management systems, was referenced by Implementing Decision (EU) 2026/2274 in October 2026.
What fines apply?
The penalties regime (Article 99) has applied since 2 August 2025, but a fine can only target an obligation that itself applies. The caps are as follows, whichever amount is higher:
| Infringement | Maximum fine |
|---|---|
| Prohibited practices (Article 5) | EUR 35 million or 7% of worldwide annual turnover |
| Operator obligations, including transparency (Article 50) | EUR 15 million or 3% |
| Incorrect, incomplete or misleading information to authorities | EUR 7.5 million or 1% |
For SMEs and start-ups, the lower amount applies (Article 99(6)). The Omnibus extends that rule to SMCs for the last two categories (new paragraph 6a) and adds the obligations in Article 25(2) and (4) to the list of fineable breaches. National authorities impose the fines; the Commission fines providers of general-purpose AI models directly (Article 101).
VerifactAI applies a versioned rule pack that reflects these dates to every declared system, and identifies which obligations apply today and which ones to prepare for 2027.
Sources
- Regulation (EU) 2024/1689, EUR-Lex
- Regulation (EU) 2026/1744 (Digital Omnibus on AI), EUR-Lex
- Article 113, AI Act Service Desk
- Article 111, AI Act Service Desk
- Article 99, AI Act Service Desk
- Implementing Decision (EU) 2026/2274 on harmonised standards, EUR-Lex
This article is general information, not legal advice.