Earlier this month, the European Commission published the final Code of Practice on the marking and labelling of AI-generated content, providing practical guidance for organisations preparing for the AI Act’s transparency obligations, which will take effect on 2 August 2026.
The voluntary Code outlines how providers and deployers of generative AI systems can comply with requirements to clearly identify AI-generated or AI-manipulated content. Under the AI Act, deepfakes and AI-generated or AI-altered text published on matters of public interest must be labelled, while users must also be informed when they are interacting with AI systems such as chatbots.
Developed by six independent experts with input from more than 180 stakeholders, the Code includes separate sections for providers and deployers. For providers, it explains how AI-generated audio, images, video, and text should be marked in a machine-readable and detectable way. For deployers, it sets out obligations to label deepfakes and certain AI-generated public-interest content where there has been no human editorial review.
The Commission has opened the Code for signatures and is encouraging providers and deployers across the EU to participate. Once approved by the Commission and the AI Board, signatories will be able to use the Code to demonstrate compliance with the relevant AI Act transparency requirements. Additional Commission guidelines will follow to clarify legal obligations and provide further practical support.
More information can be found on the European Commission’s website.
