European Commission Publishes Final Code of Practice on Marking and Labelling AI-Generated Content
On June 10, 2026, the European Commission published the final Code of Practice (the "Code") on marking and labelling AI-generated content, providing a practical compliance pathway for providers and deployers of generative artificial intelligence ("AI") systems before the EU AI Act's transparency obligations take effect on August 2, 2026.
The Code is voluntary, but it will likely become an important reference point for demonstrating compliance with Article 50 of the AI Act.
The Code addresses transparency risks associated with synthetic and manipulated content created using AI, including the risk that such content could deceive people or erode trust in public information. Starting August 2, 2026, the AI Act will require, among other things, notice when users interact with an AI system, machine-readable marking for AI-generated or AI-manipulated outputs, and clear labels for deepfakes and certain AI-generated or AI-manipulated text published on matters of public interest labelled as such (see our previous Commentary on the Code's first draft here).
What the Code Does
The Code is divided into two sections.
- Providers: The provider section focuses on Article 50(2) obligations for generative AI systems. Providers must ensure that AI-generated or AI-manipulated audio, image, video, and text outputs are marked in a machine-readable format and detectable as artificially generated or manipulated. A key development since the first draft is the Code's recognition that, under the current state of the art, no single marking technique can fully meet the AI Act's requirements of effectiveness, interoperability, robustness, and reliability. The Code therefore sets out a multi-layered approach, requiring signatories to use at least two layers of machine-readable marking where necessary, such as combinations of metadata, watermarks, or other technical measures.
The final Code also explains how AI-generated content can be detected. Providers are expected to offer detection mechanisms that allow deployers, users, competent authorities, researchers, civil society, media organizations, and other legitimate parties to verify whether content has been generated or manipulated by AI. This is intended to support transparency across the AI value chain, not just when the content is created.
- Deployers: The deployer section focuses on Article 50(4) obligations. Deployers must clearly label deepfakes and AI-generated or AI-manipulated text published to inform the public on matters of public interest, unless the text has undergone human review or editorial control and editorial responsibility has been assumed.
The Code includes guidance on how and where to display these labels, including through the use of the publicly available EU icon or equivalent labels. Where visual labels are not possible, such as audio-only content, the Code contemplates alternatives, including audio disclaimers.
Next Steps
The Code, published on June 10, 2026, is open for signature and is being assessed for adequacy by the Commission and the AI Board.
If assessed positively, signatories may rely on the Code's measures to demonstrate compliance with the AI Act's marking, detection, and labelling rules, reducing administrative burden and increasing legal certainty across Member States. Providers and deployers that choose other compliance routes will need to demonstrate the adequacy of those measures to market surveillance authorities.
The Commission will also issue guidelines clarifying the scope of Article 50 and addressing matters not covered by the Code.
