What Did Regulation (EU) 2026/1744 Change in the AI Act?

What is Regulation (EU) 2026/1744?

Regulation (EU) 2026/1744 is the Digital Omnibus on AI, an amending regulation that changes Regulation (EU) 2024/1689, the EU AI Act. It was signed on 8 July 2026, published in the Official Journal on 24 July 2026 and entered into force on 27 July 2026, the third day after publication. Because it amends the AI Act itself, the authoritative text for planning is now the consolidated version of Regulation (EU) 2024/1689 as amended, not the original 2024 text alone. Older summaries, Service Desk articles that reproduce the original text and graphics built before July 2026 can therefore carry superseded dates. Whenever a plan cites a date or a provision, the citation should name Regulation (EU) 2026/1744 where the amended text is the source, and the consolidated text should be checked before the date is relied on for a release or contract decision.

Which application dates did it move?

The most consequential change is to Article 113. Chapter III Sections 1 to 3, the high-risk classification rules, requirements and obligations, now apply from 2 December 2027 for AI systems classified as high-risk under Article 6(2) and Annex III, and from 2 August 2028 for systems classified under Article 6(1) and Annex I. Before the amendment, those dates were 2 August 2026 and 2 August 2027 respectively. The change is a deferral of application, not a removal of the requirements: classification work, documentation, testing and oversight design still need to be built, and the later dates do not suspend duties that were already applying. Plans that still show the original high-risk dates should be corrected to the amended Article 113 dates, with the amendment named, so that a reader can see why the calendar changed and which source supports the new date.

For the broader regulatory framework, see EU AI Act timeline.

What new prohibitions did it add?

Regulation (EU) 2026/1744 added two prohibitions to Article 5, numbered Article 5(1)(ba) and Article 5(1)(bb), which apply from 2 December 2026. In general terms, they concern AI systems used to generate or manipulate non-consensual intimate material and material within Directive 2011/93/EU on child sexual abuse material. This guide describes them only in those general terms, as the verified date list requires: it does not reproduce graphic detail, does not give examples involving identifiable people and does not link to offending material. Organisations should update prohibited-practice screening so these additions are tested alongside the prohibitions that have applied since 2 February 2025, examining intended purpose, actual use, control over generation and safeguards. Because prohibition analysis is use-case specific, a broad product label is not enough to decide whether either new prohibition is engaged.

What else did the Omnibus change?

Beyond dates and prohibitions, Regulation (EU) 2026/1744 made targeted changes that planning teams should not overlook. It rewrote the AI literacy provision in Article 4, and it added provision for special-category data to be used for bias detection in defined circumstances, which should be checked in the consolidated text before any reliance. It simplified aspects of registration, including the information registered for systems where a provider concludes under Article 6(3) that an Annex III system is not high-risk, and it extended simplified treatment beyond small and medium-sized enterprises to small mid-caps in defined respects. It also addressed governance, including centralised oversight of specified general-purpose AI based systems, and expanded sandbox provision. Each of those changes is provision-specific, so a team should work from the consolidated article concerned rather than assuming a general simplification applies to its system.

What did the Omnibus not change?

The Omnibus did not move the duties that were already applying. AI literacy and the original prohibitions have applied since 2 February 2025, general-purpose AI model obligations since 2 August 2025, and Article 50 transparency duties since 2 August 2026, with enforcement powers starting for applicable rules on that last date. It did not create a general grace period, and it did not remove the need for classification, documentation or oversight work before the later high-risk dates. It also did not settle the broader Digital Omnibus package covering other areas of digital law, which is distinct from Regulation (EU) 2026/1744 and still in the legislative process; the two should never be conflated in a plan or a citation. Finally, the Omnibus did not by itself publish harmonised standards or designate notified bodies, so readiness claims about standards and bodies need their own dated primary source.

Continue with how AI Act transitions work for systems already on the market.

What does this guide not decide?

This guide does not give a complete account of every amendment in Regulation (EU) 2026/1744, does not classify any system under the amended text and does not advise any reader on what they must do for a specific system. The amending regulation touches many provisions, and a provision-level question should be answered from the consolidated article, the relevant recital and current official guidance for that provision. The guide also does not predict further simplification proposals, which remain expected only until enacted text is verified in the Official Journal. Where a decision turns on an amended provision, confirm the current consolidated wording with a competent authority or legal adviser, and record the version of the text checked and the date of the check.

Frequently asked questions

Continue through this AI Act cluster

• EU AI Act timeline
• how AI Act transitions work for systems already on the market
• which AI Act rules to check now in late 2026
• what changes on 2 December 2027
• EU database, registration and post-market monitoring
• EU AI Act overview