Labelling AI Content in the EU: What Marketers Must Do Since August 2026

Since 2 August 2026, the transparency rules of the EU AI Act apply. Article 50 splits the work in two: the companies that build generative AI tools must mark their output in a machine-readable way, and the people who use those tools to publish content, called deployers, must tell the audience in certain cases. Brands and agencies publishing AI-made material are deployers. Here is what that means in practice. This is a practical summary, not legal advice.

What deployers must disclose

• Deepfakes: image, audio or video content generated or manipulated by AI that resembles real people, objects, places or events and could falsely appear real.
• AI-generated or manipulated text published to inform the public on matters of public interest, unless it has gone through human review or editorial control and someone holds editorial responsibility for it.

For evidently artistic, creative, satirical or fictional work, the Act allows the disclosure to be limited, so that it does not spoil the work itself. A fantasy dragon in a clearly fictional ad is a different case from a realistic “customer testimonial” of a person who never existed.

What it means for everyday marketing

• Realistic people, places or events made with AI: label them, clearly and where the viewer sees the content, not in a footer or the terms.
• Fake testimonials and invented “real customers”: do not make them. Beyond labelling, they are misleading advertising under consumer law anyway.
• AI-assisted blog posts and articles: keep a human editor responsible for every published piece. That is good practice, and it is what the text rules hinge on.
• Old material: content generated and published before 2 August 2026 does not need retroactive labels.

The Code of Practice

In June 2026 the European Commission published a voluntary Code of Practice on transparency of AI-generated content, with one section for tool providers (marking and detection) and one for deployers (disclosure and labelling). Signing is voluntary, but it is a useful reference for what “clear disclosure” looks like.

A simple routine we follow

1. Ask one question per asset: could someone take this for a real person, place or event? If yes, label it.
2. Label visibly, in the image or video or right next to it, in the language of the audience.
3. Keep the receipts: prompts, reference images and the tool used for every delivered shot.
4. Never generate an identifiable real person without their written consent.
5. Keep a named editor on every text that goes out.

On our own site, the AI-generated examples in the Influencer Lab carry a label, and we apply the same routine to client work. If you want to use AI in your marketing without guessing about the rules, our AI advising and AI video production teams can set it up with you. For what AI video is good at in the first place, read VFX vs AI video.