Less than a month remains until one of the most important deadlines in AI regulation worldwide: on 2 December 2026, the grace period ends for marking AI-generated content under Europe's AI Act. From that date, providers of generative AI systems that were already on the European market before 2 August 2026 must mark the content they produce — audio, image, video or text — in a machine-readable format, so it can be detected as artificial.
The duty sits in Article 50(2) of Regulation (EU) 2024/1689, the AI Act. In practice, the rule has applied since 2 August 2026 to systems launched from that date. The 2 December deadline is only the end of the four-month grace period granted to older systems, set by the Digital Omnibus on AI, Regulation (EU) 2026/1744.
What is already in force
Article 50 of the AI Act brings together four transparency duties, in force since 2 August 2026: providers must tell users when they are interacting with an AI system; providers must mark synthetic content in a machine-readable way; anyone using emotion recognition or biometric categorisation must inform the people affected; and anyone publishing deepfakes — or AI-generated text on matters of public interest — must make the artificial origin clear, unless the text went through human review with editorial responsibility.
What changes on 2 December
Besides the end of the marking grace period, two new prohibitions added by the Digital Omnibus take effect the same day: AI systems that generate non-consensual intimate images of identifiable people and systems that generate child sexual abuse material become prohibited practices under the AI Act.
Fines for breaching the transparency duties can reach €15 million or 3% of a company's worldwide annual turnover. The European Commission published interpretive guidelines on Article 50 in July 2026 and, in June, the Code of Practice on transparency of AI-generated content, recognised as a suitable reference for demonstrating compliance.
For companies and developers operating in Europe, the message is direct: the 2 December deadline is not the start of the transparency rules — it is the end of the last grace period. Anyone who still does not mark the content their systems generate needs to comply now.
The measure comes as other governments also tighten AI oversight: New York opened registration under the RAISE Act, placing OpenAI and Anthropic under official supervision.
Featured image: flag of Europe. Credit: Wikimedia Commons (public domain); flag design by the Council of Europe.
Fontes
- Regulation AI — “EU AI Act Article 50: In Force Since 2 August 2026”
- AdSovereign — “EU AI Act and Advertising: What Applies to Ad Tech”
- Medium — “EU AI Act: 5 Things to Check Before 2 December 2026”
Transparency: this article was produced with the help of artificial intelligence, based on public sources on the internet, including specialist analyses of the European AI Act. We recommend consulting the original sources listed above to verify the information.
By Muse
