The EU's AI Act introduces New Labelling Obligations

Important note: This article is intended for general information purposes only and does not constitute legal advice.

According to the European Commission itself, the "Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence" is the world's first comprehensive legal framework for artificial intelligence. With these rules, regulators aim to promote trustworthy AI across Europe. The so-called AI Act entered into force on 2 August 2026.

Although the AI Act is not primarily aimed at businesses in the packaging value chain, Article 50 – entitled "Transparency obligations for providers and deployers of certain AI systems" – sets out requirements that every publicly active company in the sector should at least review for itself.

Generated or Edited Content

This concerns the new labelling obligations for content that has been generated or altered with the help of AI. Going forward, such media content will in many cases need to be clearly marked as such. The obligations distinguish between providers and deployers. Put simply, providers create the technical foundations – in other words, they develop AI systems – whereas deployers act as users of those systems.

As users of AI tools, companies are most likely to encounter the new transparency obligations when creating materials for advertising and marketing, social media or press work, product visualisations, or corporate presentations and videos.

Transparency Obligations

The most important transparency obligations for deployers include the following:

• Labelling of content: In this respect, the Regulation initially refers only to text, not to other media formats. The provision reads as follows: "Deployers of an AI system that generates or manipulates text which is published with the purpose of informing the public on matters of public interest shall disclose that the text has been artificially generated or manipulated." (Paragraph 4)
• Across media formats, however, the labelling obligation for deployers applies to deepfakes. Deepfakes are realistic-looking image, audio or video content depicting existing persons, objects, places or events that could be mistakenly perceived as genuine. Paragraph 4 states: "Deployers of an AI system that generates or manipulates image, audio or video content constituting a deep fake, shall disclose that the content has been artificially generated or manipulated." (Paragraph 4)

For all labelling obligations, the timing and form of the required notice are equally important. The information must be provided to users clearly, unambiguously and in an accessible manner no later than the point of their first interaction with, or exposure to, the content in question.

Exemptions

• With regard to created or edited media content, the labelling obligation does not apply "…where the AI-generated content has undergone a process of human review or editorial control and where a natural or legal person holds editorial responsibility for the publication of the content." (Paragraph 4)
• There are also exemptions where "the content forms part of an evidently artistic, creative, satirical, fictional or analogous work or programme". In such cases, "the transparency obligations set out in this paragraph are limited to disclosure of the existence of such generated or manipulated content in an appropriate manner that does not hamper the display or enjoyment of the work." (Paragraph 4) 

What About Internal Communication Formats?

As the obligations set out in the Regulation are largely worded in an addressee-neutral manner, they apply in principle to any natural person – including, therefore, to internal company communications. At the same time, however, Paragraph 4 refers to AI-generated or AI-manipulated text "which is published with the purpose of informing the public on matters of public interest". This suggests that internal newsletters, intranet articles or reports are exempt, provided they are not publicly accessible or do not have any public effect.

Options for Action

As the AI Act entered into force on 2 August 2026, companies should promptly review where and to what extent they use AI-generated or AI-edited image, audio or video content, in order to implement the necessary labelling. It would, for example, be advisable to map out the use of AI-generated image, audio and video content within the company and to clarify responsibilities and approval processes for the content concerned. The European Commission provides suitable labelling notices (icons) and issues recommendations on how they should be placed (see below). External agencies and service providers should also be made aware of the new requirements.

Recommendations, Code of Practice and Materials from the European Commission

In addition to the legal text itself, the European Commission also provides a code of practice on the transparency of AI-generated content, as referred to in Article 50, paragraphs 2, 4 and 5 of the AI Act.

The code sets out commitments and measures that deployers of AI systems generating deepfakes and certain types of text can adopt to demonstrate compliance. It also contains the aforementioned guidance on placement for fulfilling the transparency obligations.

The code of practice is available here.

The AI Act itself can be found here. 

The EU also provides icons for labelling AI-generated content.

Author: Christian Nink, Freelance Journalist