What the Digital Omnibus changed (in force since 27 July 2026)

Partly delayed, and now settled law. Regulation (EU) 2026/1744 of 8 July 2026, the Digital Omnibus on AI, amends Regulation (EU) 2024/1689, the EU AI Act. Parliament voted on 16 June 2026 and the Council on 29 June 2026. It was published in the Official Journal (OJ L) on 24 July 2026 and has been in force since 27 July 2026, the third day after publication (Article 4 of Regulation 2026/1744).

So "is the AI Act delayed?" has a precise answer. The high-risk rules moved later. Other rules did not move, and the Omnibus added some new dates. The Omnibus also changed what some articles require. This page sets out each change and the date that now applies.

---

Short answer: the high-risk rules moved, the rest did not

What moved

A new Article 113 point (c) defers Chapter III Sections 1–3 (all except Article 6(5)):

• Stand-alone high-risk systems classified through Article 6(2) and Annex III: from 2 August 2026 to 2 December 2027.
• Product-embedded high-risk systems classified through Article 6(1) and Annex I: from 2 August 2027 to 2 August 2028.

These are fixed calendar dates. They apply whether or not harmonised standards are ready by then, so plan back from them.

Section 5 (Arts. 40–49) is not named in the Art. 113(c) deferral, but it has no practical effect until the deferred Sections 1–3 apply.

What did not move

• The Article 5 prohibitions have applied since 2 February 2025.
• The obligations for general-purpose AI models under Articles 51–55 have applied since 2 August 2025.
• Article 50, including the Article 50(2) marking duty, has applied since 2 August 2026 (see below).

---

Article 50 has applied since 2 August 2026

The second paragraph of Article 113 ("shall apply from 2 August 2026") was not amended, and Article 50(2) was not amended or deferred. The transparency duties in Article 50, including the Article 50(2) duty to mark AI-generated content, have applied since 2 August 2026.

The new Article 111(4) gives a transition only to generative systems placed on the market before 2 August 2026. Those systems have until 2 December 2026. A generative system placed on the market on or after 2 August 2026 has no transition.

---

Two new Article 5 bans from 2 December 2026

The Omnibus adds two prohibited practices to Article 5. Both apply from 2 December 2026:

• Point (ba): AI systems that generate non-consensual intimate imagery of an identifiable person. This is broader than "nudifier apps".
• Point (bb): AI systems that generate child sexual abuse material (CSAM).

The new Article 5(1a) and (1b) set the scope of both bans and the safeguards conditions. Read them before you decide that a generative system falls outside the ban.

---

The Omnibus also changed substance

The Omnibus does more than move dates. Three substantive changes matter for most programmes:

1. Article 4 (AI literacy) was replaced. The new text is a softer duty than the original.
2. Article 6(1a) and (1b) were inserted. They narrow the Annex I safety-component test, which decides when AI in a regulated product is high-risk. Check each product function against the amended text.
3. Annex VIII Section B, points 7 and 9, were deleted.

---

Systems already on the market: Article 111(2)

The new Article 111(2) limits how the high-risk rules reach legacy systems. A high-risk system already on the market is caught only if its design changes significantly. High-risk systems intended to be used by public authorities are the exception: they must comply by 2 August 2030.

---

The Article 6 classification guidelines are still in draft

Article 6(5) is the one part of Sections 1–3 that was not deferred. The Commission published draft guidelines on high-risk classification. The consultation opened on 19 May 2026, was extended from 23 June, and closed on 23 July 2026. The Commission says it will adopt the final guidelines by the end of 2026. As of 23 October 2026 we found no final guidelines. Until they appear, the draft is the best indication of the Commission's reading, but it is not binding.

---

Deadlines at a glance

Source: Regulation (EU) 2026/1744 of 8 July 2026, OJ L 24.7.2026. For every date in the Act, see the full EU AI Act deadline timeline.

Section 5 (Arts. 40–49) is not named in the Art. 113(c) deferral, but it has no practical effect until the deferred Sections 1–3 apply.

---

What to do in the runway

The high-risk deferral gives you time to prepare. Do not read it as a reason to wait.

1. Close the live gaps first. Article 5 and Article 50 apply today. If a generative system you put on the market before 2 August 2026 does not yet mark its output, you have until 2 December 2026.
2. Check the two new bans against your generative systems before 2 December 2026.
3. Inventory and classify every AI system. Annex III systems (Article 6(2)) and Annex I products (Article 6(1)) carry different dates. Re-check Annex I products against Article 6(1a) and (1b).
4. Date-stamp classifications against the draft guidelines and re-run borderline systems when the final text appears.
5. Start the long work now: the Article 9 risk management system, Article 10 data governance and the Article 11 / Annex IV technical documentation, ahead of the Article 43 conformity assessment.

See how enforcement is phasing in and the EU AI Act implementation timeline for the surrounding sequence.

---

Penalties

The fine ceilings in Article 99 for operators:

Each tier takes the higher of the fixed amount and the percentage for most undertakings. For SMEs and start-ups, Article 99(6) caps each fine at the lower of the two. See the penalty tiers under Article 99.

---

How Confir helps

Confir's rule set reflects Reg. 2026/1744 as published. Its classification logic is deterministic and rule-based. It asks plain-English questions about what each AI system does and derives the risk tier under Articles 5 and 6. It separates Annex III stand-alone systems (Article 6(2), from 2 December 2027) from Annex I product-embedded systems (Article 6(1), from 2 August 2028). The same answers always produce the same result.

From that classification, Confir shows which obligations apply today (Article 5, Article 50) and which sit on the deferred high-risk runway, so you deal with current gaps first. For high-risk systems it builds the Article 11 / Annex IV technical documentation and the Article 47 / Annex V EU Declaration of Conformity. Every classification decision and its rationale is time-stamped and locked against edits.

---

Frequently asked questions

Is the EU AI Act delayed?

Partly. Regulation (EU) 2026/1744 (the Digital Omnibus), in force since 27 July 2026, moves the high-risk rules in Chapter III Sections 1–3 to 2 December 2027 for Annex III systems (Art. 6(2)) and 2 August 2028 for Annex I products (Art. 6(1)). The Article 5 prohibitions, the GPAI rules and the Article 50 transparency duties were not deferred.

Is the Digital Omnibus law yet?

Yes. Regulation (EU) 2026/1744 of 8 July 2026 was published in the Official Journal on 24 July 2026 and entered into force on 27 July 2026, the third day after publication.

When do the Article 50 transparency and marking duties apply?

Article 50, including the Article 50(2) marking duty, has applied since 2 August 2026. Only generative systems placed on the market before 2 August 2026 have until 2 December 2026 (Art. 111(4)).

What new prohibitions did the Omnibus add?

Two new Article 5 bans apply from 2 December 2026: point (ba), AI systems that generate non-consensual intimate imagery of an identifiable person, and point (bb), child sexual abuse material. Article 5(1a) and (1b) set their scope and safeguards conditions.

Did the Omnibus change any obligations, or only dates?

It changed substance too. Article 4 on AI literacy was replaced with a softer duty, Article 6(1a) and (1b) were inserted to narrow the Annex I safety-component test, and Annex VIII Section B points 7 and 9 were deleted.

What happens to high-risk systems already on the market?

Under the new Article 111(2), legacy high-risk systems are caught only if their design changes significantly. High-risk systems intended to be used by public authorities must comply by 2 August 2030.

Are the Article 6 classification guidelines final?

Not yet. The consultation on the draft guidelines ran from 19 May to 23 July 2026 (extended from 23 June). The Commission says it will adopt the final guidelines by the end of 2026; none had been published as of 23 October 2026.

---

Related guides

• The Digital Omnibus and the 2027 high-risk deadline
• The full EU AI Act deadline timeline
• The EU AI Act implementation timeline
• How enforcement is phasing in
• The Article 5 prohibitions
• The penalty tiers under Article 99