Compliance and Europe
The EU AI Act and marketing content: obligations for brands
Quick answer
Under the EU AI Act, Regulation (EU) 2024/1689, Article 50 transparency obligations apply from 2 August 2026. Brands using AI to create realistic images, audio or video of people, places or events must disclose that the content is AI-generated. The Digital Omnibus, Regulation (EU) 2026/1744, kept this date and gave providers until 2 December 2026 for watermarking.
Key takeaways
- Article 50 of Regulation (EU) 2024/1689 has applied since 2 August 2026; the Digital Omnibus on AI, Regulation (EU) 2026/1744, did not postpone it.
- Most brands act as deployers, so their core duty is to disclose AI-generated or manipulated images, audio and video that resemble real people, objects, places or events.
- Generative AI providers whose systems were on the market before 2 August 2026 have until 2 December 2026 to apply machine-readable marking under Article 50(2).
- The Commission published the final Code of Practice on marking and labelling AI-generated content on 10 June 2026 and Article 50 guidelines on 20 July 2026.
- Breaches of Article 50 can be fined up to EUR 15 million or 3% of worldwide annual turnover, whichever is higher.
The EU AI Act now governs AI-generated marketing content: since 2 August 2026, brands must disclose realistic AI-generated or manipulated images, audio and video, and the tools they use must mark outputs in machine-readable form. Most brands are deployers, so their obligations are disclosure, documentation and staff literacy. The DAM is where those obligations become operational, because it records which assets are synthetic and carries the labels and provenance data to every channel.
The EU AI Act timeline after the Digital Omnibus
The AI Act, Regulation (EU) 2024/1689, entered into force on 1 August 2024 and applies in stages. Prohibited practices and the Article 4 AI literacy duty have applied since 2 February 2025, and general-purpose AI model rules since 2 August 2025. The Digital Omnibus on AI, Regulation (EU) 2026/1744, published on 24 July 2026 and in force since 27 July 2026, moved high-risk obligations to 2 December 2027 (stand-alone systems) and 2 August 2028 (AI in regulated products). It left the Article 50 transparency date of 2 August 2026 unchanged, adding only a watermarking grace period until 2 December 2026 for generative systems placed on the market before 2 August 2026. Marketing teams should therefore treat transparency as live today.
Article 50 obligations by role
Article 50 splits duties between providers, who build or place AI systems on the market, and deployers, who use them. A brand becomes a provider when it builds its own customer-facing chatbot or offers an in-house image generator to partners.
| Marketing use case | Brand's role | Obligation | Applies from |
|---|---|---|---|
| AI-generated campaign image of a realistic person or place | Deployer | Disclose generation or manipulation (Article 50(4)) | 2 August 2026 |
| Synthetic voice or video presenter in an advert | Deployer | Disclose (Article 50(4)) | 2 August 2026 |
| Own chatbot on the brand website | Provider | Inform users they interact with AI (Article 50(1)) | 2 August 2026 |
| Own generative tool offered to partners | Provider | Machine-readable marking (Article 50(2)) | 2 August 2026, or 2 December 2026 if on the market earlier |
| AI-drafted product copy reviewed by editors | Deployer | No Article 50 disclosure | Not applicable |
Disclosure must be clear and made at the latest at first exposure. For evidently artistic, creative, satirical or fictional work, the disclosure may be adapted so it does not hamper enjoyment of the work.
The Code of Practice and Commission guidelines
The Commission published the final Code of Practice on marking and labelling AI-generated content on 10 June 2026 and guidelines on Article 50 on 20 July 2026. The Code has a provider section on machine-readable marking, recommending at least two layers such as metadata and watermarks, and a deployer section on labelling deepfakes, where signatories can use a common EU icon. The Code is voluntary, and organisations that follow other approaches must show market surveillance authorities that their measures are adequate. Breaches of Article 50 can be fined up to EUR 15 million or 3% of worldwide annual turnover.
What the AI Act means for DAM and content operations
The AI Act turns provenance into required asset metadata. Each asset should record whether it is AI-generated or AI-edited, the tool used, the prompt owner and the disclosure status. Content Credentials based on the C2PA standard give a machine-readable layer that survives distribution when export profiles preserve metadata, as explained in Content Credentials and C2PA. Adobe, owner of Adobe Experience Manager Assets, launched the Content Authenticity Initiative in 2019, and platforms such as Bynder and Aprimo are worth testing on whether AI flags and credentials pass through to delivery URLs. The wider use case picture sits in generative AI in DAM.
Governance: who owns AI disclosure decisions
Disclosure decisions need a named owner, a written policy and an audit trail. Brand or legal teams should define which asset categories count as deepfakes, which labels apply per channel and who signs off borderline cases, following a structure like the AI governance framework for brand assets. Product descriptions generated in PIM follow separate rules, covered in AI-generated product descriptions. We estimate that a mid-market brand can set up the metadata fields, labels and approval step in 4 to 8 weeks.
What to do next
- Inventory every AI tool used for marketing content and classify your role as provider or deployer per tool.
- Add AI-generated, AI-edited, tool and disclosure status fields to the DAM metadata model.
- Adopt a label standard per channel, using the EU icon from the Code of Practice where it fits.
- Configure export and CDN profiles to preserve C2PA Content Credentials and embedded metadata.
- Record AI literacy training for staff who create or approve AI-generated assets.
Frequently asked questions
Do brands have to label AI-generated images in advertising?
Yes, when the image is a deepfake in the AI Act sense: AI-generated or manipulated content that resembles existing people, objects, places or events and would falsely appear genuine. For evidently creative or fictional work, disclosure can be made in a way that does not hamper the display. Stylised illustrations that nobody would mistake for reality generally fall outside the deepfake definition.
Do AI-written product descriptions need a disclosure under the AI Act?
Usually no. The Article 50(4) text obligation covers AI-generated text published to inform the public on matters of public interest, and it falls away when a human reviews the text and a person or company holds editorial responsibility. Product copy is commercial content, though consumer law on misleading claims still applies.
Did the Digital Omnibus delay the AI Act transparency rules?
No. Regulation (EU) 2026/1744, in force since 27 July 2026, postponed high-risk obligations to 2 December 2027 and 2 August 2028. Article 50 still applies from 2 August 2026, with only a grace period until 2 December 2026 for watermarking by generative AI systems already on the market.
Sources
- EUR-Lex: Regulation (EU) 2026/1744 (Digital Omnibus on AI)
- Law and Technology: Digital Omnibus on AI published in the Official Journal
- Jones Day: European Commission publishes final Code of Practice on marking and labelling AI-generated content
- European Commission: Guidelines on transparency obligations for providers and deployers of certain AI systems