The European Commission has published the first draft of a Code of Practice on Transparency of AI-Generated Content, developed as part of the implementation of the EU Artificial Intelligence Act. The draft outlines voluntary commitments and practical measures intended to support providers and deployers of generative AI systems in meeting transparency obligations under Article 50 of the AI Act, ahead of the rules becoming applicable in August 2026.
The Code is designed to help ensure that AI-generated and AI-manipulated content — including text, images, audio and video — can be reliably identified and disclosed. It proposes guidance on marking and detection techniques (such as machine-readable metadata and other signals) for providers, and on clear labelling for deployers, such as when publishing deepfakes or synthetic content in professional contexts. The first draft results from a collaborative process involving industry, academia and civil society, and will be refined further based on stakeholder feedback ahead of an expected final version in 2026.
