The Digital Omnibus on AI, Regulation (EU) 2026/1744, enters into force today, 27 July 2026. It amends the EU AI Act: obligations for high-risk AI systems move to 2 December 2027 and 2 August 2028, generative AI systems already on the market get until 2 December 2026 to mark their outputs, and AI tools for non-consensual intimate deepfakes become prohibited. What it does not touch matters as much for IP owners: the copyright duties of general-purpose AI (GPAI) model providers, and the 2 August 2026 date from which the Commission can fine them, stay as they were.
Key takeaways
- High-risk obligations now apply from 2 December 2027 (Annex III uses) and 2 August 2028 (AI in regulated products).
- Providers of generative systems placed on the market before 2 August 2026 must comply with machine-readable marking (Article 50(2)) by 2 December 2026.
- New prohibitions on AI systems that generate intimate images of identifiable people without consent, or child sexual abuse material, apply from 2 December 2026.
- The text does not amend Article 53: the copyright policy and the training content summary remain mandatory for GPAI providers.
What is the Digital Omnibus on AI?
It is a targeted amendment to the AI Act (Regulation (EU) 2024/1689), adopted to ease implementation after delays in harmonised standards and national authorities. The European Parliament adopted its position on 16 June 2026, the Council approved it on 29 June, and the act was signed on 8 July and published in the Official Journal on 24 July 2026. It enters into force on the third day after publication, which is today, as Lewis Silkin notes (27 July 2026). The AI Act’s general date of application, 2 August 2026, is unchanged.
Which AI Act dates has the Digital Omnibus changed?
| Obligation | Original AI Act | After the Omnibus |
|---|---|---|
| High-risk systems listed in Annex III (employment, credit, education…) | 2 August 2026 | 2 December 2027 |
| High-risk AI in products covered by Annex I | 2 August 2027 | 2 August 2028 |
| Machine-readable marking (Art. 50(2)) for systems already on the market before 2 August 2026 | 2 August 2026 | 2 December 2026 |
| Prohibition of intimate deepfakes and child sexual abuse material (new Art. 5(1)(ba) and (bb)) | Not provided for | 2 December 2026 |
| GPAI provider obligations (Art. 53) | 2 August 2025; 2 August 2027 for models already on the market | Unchanged |
| Commission fines on GPAI providers (Art. 101) | 2 August 2026 | Unchanged |
Other changes include a softer AI literacy duty (Article 4 now asks providers and deployers to “take measures to support” literacy), a new legal basis for processing sensitive data to detect bias (Article 4a) and lighter regimes for small mid-cap enterprises.
What did not change for copyright and GPAI models?
The Omnibus does not amend Article 53 of the AI Act, and the word “copyright” does not appear in it. GPAI providers must still keep a policy to comply with EU copyright law, including honouring text and data mining opt-outs under Article 4(3) of the DSM Directive, and publish a sufficiently detailed summary of training content. Recital 106 of the AI Act still applies this wherever the training took place.
Two governance changes do affect GPAI providers:
- Codes of practice no longer need approval by implementing act. The Commission assesses whether a code is adequate and publishes that assessment (new Article 56(6)).
- The AI Office gains exclusive competence over AI systems built on a GPAI model by the same provider or group, and over AI in very large online platforms and search engines (Article 75), with powers to inspect premises, copy records and seal premises (new Article 75a).
The new prohibitions on intimate deepfakes
From 2 December 2026, Article 5 bans placing on the market, putting into service or using AI systems that generate realistic intimate or sexually explicit material of an identifiable person without explicit consent, or child sexual abuse material. For providers the ban bites where that output is the intended purpose, or is reasonably foreseeable without adequate safeguards. Breaches of Article 5 carry fines of up to EUR 35 million or 7% of worldwide turnover (Article 99(3)). For brands and talent, this adds a regulatory route alongside image rights and personality claims.
What this means for your business
- Rights holders: do not wait. The copyright policy and training summary are enforceable obligations, and the Commission can fine GPAI providers from 2 August 2026. Keep your text and data mining opt-outs in machine-readable form and document them.
- AI providers outside the EU, including in Latin America: the territorial scope is unchanged. If your model or system reaches the EU market, check marking by 2 December 2026 and your copyright policy now.
- Companies deploying high-risk AI: use the extra time to build documentation, not to stop the project.
If you need the copyright and regulatory sides aligned, our team for AI and digital asset IP protection can review your opt-outs, licences and AI vendor terms, and AI Act compliance is covered within the same group.
Where companies get the Omnibus wrong
- Reading “simplification” as a general delay. Only certain dates moved; GPAI and copyright obligations did not.
- Assuming the marking grace period covers new products. It applies only to systems placed on the market before 2 August 2026.
- Treating the Code of Practice as optional paperwork. It remains the main way to show compliance with Article 53, now assessed by the Commission.
- Handling AI regulation and IP in separate silos, so that opt-outs, licences and transparency notices contradict each other.
Frequently asked questions
When does the Digital Omnibus on AI apply?
Regulation (EU) 2026/1744 was published in the Official Journal on 24 July 2026 and enters into force on 27 July 2026. Its provisions mostly amend AI Act dates: high-risk obligations apply from 2 December 2027 or 2 August 2028, and the new prohibitions and the marking deadline for existing systems from 2 December 2026.
Does the Omnibus change the copyright obligations of GPAI providers?
No. Article 53 of the AI Act is not amended. Providers must keep a copyright compliance policy that respects text and data mining opt-outs and publish a training content summary. The change is procedural: codes of practice are now assessed for adequacy by the Commission instead of being approved by implementing act.
Can IP Global Guard help rights holders and AI companies with these changes?
Yes. We review opt-outs, licences and contracts with AI providers, prepare and coordinate enforcement across Europe, Latin America and Africa through qualified local correspondents where needed, and work with the group’s AI Act team so that IP and regulatory positions match, all from a single point of contact.
How IP Global Guard can help after the Omnibus
The Omnibus moves some deadlines but leaves the copyright core of the AI Act intact. IP Global Guard, the IP services line of META Channel Corporation Limited, handles AI, copyright and licensing matters with one strategy and one billing relationship across more than 25 jurisdictions in Europe, Latin America and Africa.
Tell us whether you are protecting content or placing AI on the EU market, and which products and countries are involved. We will map what applies now, what moved and what to prepare first. Speak to our AI and IP team.
This article is general information, not legal advice, and reflects the position at its date of publication.
