
Code of Practice on AI Content Marking
EU voluntary code operationalising Article 50 transparency rules for marking AI-generated content.
Last refreshed: 7 July 2026
Why has no broadcaster signed the EU's AI content-marking code?
Timeline for Code of Practice on AI Content Marking
Mentioned in: Brussels stalls its own AI-label code
Media's AI PivotWhat is the EU Code of Practice on AI content marking?
Why hasn't any broadcaster signed the EU's AI labelling code?
Is the EU Code of Practice on AI content marking the same as the GPAI Code of Practice?
Background
The European Commission confirmed on 7 July 2026 that the Code of Practice remains under an adequacy assessment by the Commission and the AI Board, with no EU broadcaster yet signed and a 22 July signatory cutoff ahead of application.
The Commission finalised the Code on 10 June 2026, a voluntary framework signatories can use to demonstrate compliance with Article 50 of the EU AI Act, which requires machine-readable marking of AI-generated content from 2 August 2026. Deployers must label AI-generated text on matters of public interest unless that text sits under human editorial responsibility, the clause broadcasters read first. It is a distinct instrument from the GPAI Code of Practice, which covers general-purpose AI model providers rather than content marking, and the two should not be conflated.
The Code sits alongside the AI Omnibus, which grandfathers AI systems already on the market to 2 December 2026, leaving new deployments to the tighter 2 August Deadline. Spotify's early adoption of the DDEX metadata standard for AI-track labelling anticipated the same rulebook. Zero signatures five weeks from application reads less as industry foot-dragging than as a regulator still certifying its own instrument.