Take part in the development of the second version of the Code of Practice on transparency in AI-generated content

The European Commission and the AI Office have published the second version of the Code of Practice on the marking and labelling of content generated or manipulated by artificial intelligence, as part of the implementation of the transparency obligations set out in the European Artificial Intelligence Regulation (AI Act). The Commission invites all interested parties to submit their contributions by 30 March 2026.

This Code of Practice aims to support providers and deployers of AI systems in implementing mechanisms to identify and signal content generated or manipulated by AI, helping ensure that users are clearly informed when interacting with generative AI systems or when encountering content produced by such systems. The second version of the document can be consulted on the European Commission portal: Second version of the Code of Practice.

The second version of the Code of Practice is structured in two sections addressing different aspects of the transparency obligations set out in the AI Act for providers and deployers of artificial intelligence systems. The first section focuses on the marking and detection of AI-generated content and is aimed at providers of generative AI systems covered by Article 50, no.2. It introduces greater flexibility and clarification of the proposed measures, including a two-layer marking approach based on the use of secure metadata and watermarks, as well as optional mechanisms for fingerprint collection and verification procedures. The second section is addressed to entities that use or deploy AI systems under Article 50, no.4, focusing on the labelling of deepfakes and textual content related to matters of public interest. It establishes guidance for the design and placement of icons, labels or notices that inform users clearly and consistently, while maintaining flexibility in implementation. The document also foresees the creation of a task force dedicated to developing a common European icon for AI-generated content, along with illustrative examples of this symbol included in the annex of the draft.

The European Commission invites all interested parties to submit feedback on this second version through the questionnaire available online. The deadline for submitting responses is 30 March 2026, at 22:00 (CET), and participation can be made here: EU Survey – Transparency Code of Practice.

The process of developing the Code of Practice follows an inclusive, transparent and iterative approach led by the AI Office, involving representatives from industry, academia, public authorities and civil society. The Code aims to operationalise the provisions of the AI Act relating to user information, the detection of deepfakes and the disclosure of artificially generated content.

The AI Act has been in force since 1 August 2024, establishing a European framework for the responsible, safe and transparent development and use of artificial intelligence.