Skip to content

Code of Practice on Transparency of AI-Generated Content

The European Commission published the final Code of Practice on Transparency of AI-Generated Content on 10 June 2026. The voluntary code supports providers and deployers preparing for the AI Act article 50 transparency duties that apply from 2 August 2026.

The code provides practical compliance measures for:

  • providers marking synthetic audio, image, video and text in a machine-readable way;
  • deployers disclosing deepfakes and AI-generated or manipulated public-interest text;
  • informing people when they interact with an AI system;
  • using the Commission’s common visual icons for AI-generated content.

Signing is voluntary and does not replace the binding AI Act. It is a compliance aid: non-signatories remain responsible for demonstrating compliance by other adequate means.

At the 30 June cutoff, the Commission and AI Board were still assessing the Code’s adequacy. It was also due to be complemented by Commission guidelines on the legal scope of article 50; the May 2026 guidelines remained a draft. A final Code and draft legal guidance are therefore distinct instruments.