EU AI Act 2026: New Labelling Rules for AI-Generated Content

AI-assisted and editorially reviewed · Responsible: Rainer Schmitt

Since 2 August 2026, new transparency requirements for AI-generated content have applied across the European Union. With the new Code of Practice on Transparency of AI-Generated Content, the EU is providing practical guidance on how businesses can implement the requirements of the AI Act. For marketing agencies, online shops, companies and content creators, this means that anyone using generative AI professionally should now review the new requirements.

Why is the EU introducing new labelling requirements?

Generative AI has made enormous progress in recent years. Images, videos, audio recordings and texts can now be created within seconds, often at a quality that makes them difficult to distinguish from authentic content.

This is precisely where the EU AI Act comes into play. Its purpose is to reduce manipulation and disinformation while strengthening trust in digital content. The transparency obligations under Article 50 of the AI Act are intended to help users recognise when content has been created or altered using AI.

Further information is available on the European Commission’s website: Code of Practice for AI-generated content .

What is the Code of Practice?

The Code of Practice on Transparency of AI-Generated Content was developed by independent experts on behalf of the European AI Office. Although the Code itself is voluntary, it provides businesses with recognised guidance for demonstrating compliance with the transparency obligations of the AI Act.

Key points to remember:

• The Code itself is voluntary.
• The transparency obligations under the AI Act are mandatory.
• Businesses that follow the Code may find it easier to demonstrate that they are meeting the legal requirements.

Which types of content must be labelled?

The new requirements distinguish between several types of AI-generated content.

1. Deepfakes

Images, videos or audio recordings that realistically depict or manipulate real people, places or events must be clearly labelled as AI-generated or AI-manipulated.

2. AI-generated texts on matters of public interest

Texts dealing with topics such as politics, health, safety or socially relevant events must be labelled when they are published automatically without human editorial review.

3. Chatbots

Users must be able to recognise that they are interacting with an AI system unless this is already obvious from the context.

Does all AI-generated text now have to be labelled?

No.

This is an important point that is often misunderstood. A blog article or product description created with the assistance of AI does not automatically have to be labelled.

Where content has undergone human review and is published under editorial responsibility, a general labelling obligation will often not apply. The decisive factor is that a person checks and takes responsibility for the content instead of publishing an unreviewed AI-generated result.

What does this mean for businesses?

For businesses, the new requirements primarily create organisational responsibilities.

Companies that regularly use AI should define clear internal processes:

• Which content is created fully or partly with AI?
• Which content is reviewed before publication?
• When is labelling required?
• Which technical methods, such as metadata or watermarks, are being used?

Marketing departments, agencies and e-commerce businesses in particular can benefit from documented workflows. These processes can help demonstrate compliance with the applicable requirements during future reviews or audits.

What role do watermarks and metadata play?

In addition to visible labels, the Code also recommends the use of technical markings, such as digital watermarks or machine-readable metadata.

These methods can help platforms and other systems recognise whether content has been generated or modified using AI. The EU also provides standardised symbols that businesses may use.

What are the implications for marketing and SEO?

In the long term, the new rules may be an advantage for reputable businesses.

Transparency builds trust. Customers increasingly want to understand how content is created, particularly when it concerns sensitive topics.

For SEO, the following principles remain important:

• High-quality content remains essential.
• Human expertise is becoming increasingly important.
• AI is a tool, but it does not replace professional review.

Search engines assess content based on its quality, usefulness and trustworthiness. Whether a text was created with AI assistance is less important than its professional quality and the value it provides to readers.

Our conclusion

The new EU Code of Practice provides greater clarity for the professional use of generative AI. Although the Code itself is voluntary, the underlying transparency obligations of the AI Act have applied since 2 August 2026. Businesses should therefore review and document their content creation processes.

Companies that use AI responsibly, review content editorially and apply labels where required are not only supporting regulatory compliance. They are also strengthening the trust of their customers.