EU AI content marking: the 2 December deadline
Generative AI tools on the EU market before 2 August 2026 must mark their output as AI made by 2 December 2026, the European Commission says.
What happened
Article 50 of the EU AI Act sets transparency rules for content made or changed by AI. Most of them have applied since 2 August 2026.
One duty has a transition. Providers of generative AI systems must mark their output in a machine readable way. For systems placed on the EU market before 2 August 2026, the AI Omnibus amendment gives them until 2 December 2026.
The Commission’s voluntary Code of Practice explains how to comply. Its first section covers the tool makers and their marking. Its second section covers deployers, the companies that publish the content: label deepfakes, and label AI generated text published to inform the public on matters of public interest. Work under human review and editorial responsibility gets special treatment.
What it means for your content
The marking duty sits with the tool makers, not with you. The effect reaches you anyway. From December, more images, video and audio from mainstream tools will carry marks that platforms can read, so expect AI labels on posts you never labelled yourself.
The deepfake label is your duty, and it already applies. If a post shows a real person, place or event that AI made or changed, label it before it goes out.
Two things to set up now. Keep a record of which tool made which asset. Name the person who reviews AI made content and signs it off, because human review is what the Code treats differently.
Source
Author
Taras Kopytko, Founder, PackCraft
Published 16 Nov 2026
We turn changes like this into your monthly content plan. You approve every piece.
Book a demoMore from Insights