The European Commission has published the final Code of Practice on Transparency of AI-Generated Content, with the aim of supporting providers and users of generative AI systems in complying with the transparency obligations set out in the AI Act.
The Code is voluntary in nature and provides practical guidance on meeting the requirements that will come into force on 2 August 2026. From that date, the AI Act will require clear labeling in cases such as deepfakes and content generated or modified by AI that is published to inform the public on matters of public interest. In addition, users must be informed when they are interacting with AI systems, such as chatbots.
The Code was developed by six independent experts with contributions from more than 180 stakeholders, including businesses, academic institutions, public organizations, and civil society groups.
The text is structured into two main sections:
- Providers of generative AI systems: Measures are outlined for labeling content created or modified by AI (audio, images, video, and text) in a manner that is technically detectable.
- Users/Deployers: The obligations for labeling deepfakes and AI-generated or AI-modified content are specified when such content is published to inform the public without human oversight or editorial review.
The European Commission is now inviting providers and users of AI systems to endorse the Code. At the same time, additional guidelines are expected to be issued to clarify the scope of the relevant obligations and support the proper implementation of the AI Act’s transparency requirements.
The Code represents another step toward strengthening transparency, accountability, and trust in the use of Artificial Intelligence within the European Union.
See more here.