New EU AI Labeling Rule: What Users and Platforms Must Know Now

The central legal foundations are found in Article 50 of Regulation (EU) 2024/1689 on transparency obligations and in Article 99 on possible sanctions. The regulation is supplemented by the European Commission's "Code of Practice on Transparency of AI-Generated Content" of 10 June 2026. This code of conduct specifies how providers of AI systems and platforms should implement the labeling in technical and organizational terms.

Legal Basis in the EU

For the major social networks, the regulation means concrete adjustments. Meta, which operates Facebook and Instagram, YouTube, and TikTok have each published transparency policies on labeling AI-generated content. Posts that are entirely or in substantial parts created with AI must accordingly be visibly marked for users. Synthetic images, audio, and video content also fall under this.

Enforcement of the new requirements lies, among others, with the European AI Office, which coordinates the oversight of AI systems at the EU level. At the national level, the responsible authorities cooperate — in Austria, for example, through the Federal Chancellery. Violations of the transparency obligations can be punished with fines under Article 99 of the regulation; the amount depends on the nature and severity of the violation.

Tasks of the European AI Office

For users, everyday scrolling is what changes most: content that was previously not identifiable as AI-generated now carries visible notices. Platforms are required not to hide or remove these labels. At the same time, substantive responsibility for the published posts remains with the respective creators.

The scope of the regulation is broad. It concerns not only classic deepfake videos such as the Trump example mentioned, but also AI-generated texts, images, and audio contributions. Even when AI is used only in a supporting role, a labeling obligation may exist if the AI portion is material to the perception of the content.

Reactions and Outlook

Observers see the regulation as an important step toward strengthening digital information integrity. Critics point out that the effectiveness of the regulation depends on consistent implementation by the platforms and on oversight by the authorities. It remains to be seen how practice develops in the coming months and what further guidelines the European Commission will publish.

For the coming months, industry representatives expect initial concrete enforcement measures. The European AI Office has announced that it will carry out random checks of compliance with the transparency obligations. At the same time, discussions with platform operators are ongoing regarding technical standards for the automatic detection and labeling of AI-generated content. An adaptation of the code of conduct to new developments is also possible.