EU updates voluntary code for labelling AI-generated content
A new EU draft code suggests using metadata and digital watermarking to identify AI-generated content across platforms, publishers, and generative AI systems in Europe.
The European Commission has released a second draft of its voluntary Code of Practice on marking and labelling AI-generated content, designed to support compliance with transparency rules under the Artificial Intelligence Act.
Published on 5 March, the updated draft reflects feedback from hundreds of stakeholders, including industry groups, academic researchers, policymakers, and civil society organisations.
Revisions follow consultations held in early 2026 as part of the broader rollout of the EU’s AI regulatory framework.
The proposed code outlines technical approaches for identifying AI-generated material. A two-layered system using secure metadata and digital watermarking is recommended, with optional fingerprinting, logging, and verification to improve detection.
Guidelines also address how platforms and publishers should label deepfakes and AI-generated text related to matters of public interest. Public feedback is open until 30 March, with the final code expected in early June before transparency rules take effect on 2 August 2026.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!