Code of practice
Code of Practice on Transparency of AI-generated Content
Voluntary code of practice in two sections: section 1 on marking and detectability of AI-generated or manipulated content by providers, section 2 on labelling deepfakes and AI-generated text by deployers. The code describes watermarks, metadata and provenance information and contains the EU icons for visible labels. Signature can be per section; the deadline for initial signatories was 27 July 2026 and as of end of July 2026 approximately 190 organisations were listed. Process: consultation September 2025, first draft 17 December 2025, second draft 3 March 2026, final 10 June 2026. The code helps demonstrate compliance with Article 50 but does not replace statutory obligations.
What this means for you
For marketing, communications and media teams, this is the most concrete description of how to demonstrably mark and label AI-generated content now that Article 50 has applied since 2 August 2026.