The European Commission has published its Code of Practice on the marking and labelling of AI-generated content, giving businesses a practical roadmap to comply with upcoming European law.
“The use of these EU icons is optional, but the labelling requirements under Article 50 AI Act are not.” – as noted in the EU Icons for labelling AI-generated content by the European Commission.
Under Article 50 of the EU AI Act, compliance becomes legally mandatory across the European Union starting August 2, 2026.
Key Requirements at a Glance
- Deepfakes & Public Text: Any AI-generated or manipulated images, audio, video, or public-interest text (lacking human editorial review) must feature clear visual disclosures.
- Interactive AI: Users must be explicitly informed whenever they are interacting with an AI system, such as a customer service chatbot.
- Machine-Readable Watermarks: AI system providers are expected to embed invisible, robust, and machine-readable data into synthetic media to ensure it can be detected across platforms.
Exceptions will be made for artistic and creative works, for public interest authorised by law, and human review or editorial control when it comes to text.
To streamline the rollout and avoid confusing user interfaces, the EU has also released a standardized set of “EU AI Icons” that companies can freely integrate into their platforms to clearly flag synthetic content.
