EU AI ActWhen AI Must Be Labeled
With the EU AI Act, new transparency requirements have been in effect since August 2, 2026.
Artificial intelligence has long been part of everyday creative work. Whether it’s text, images, videos, or chatbots, many creatives now use AI as a matter of course in their work. With the EU AI Act, new transparency requirements have been in effect since August 2, 2026. However, this does not mean that all AI-generated content will have to be labeled in the future.
Designers, agencies, content creators, and media companies in particular should take a look at the new rules. After all, their primary purpose is to foster greater transparency and trust in the use of AI. People should be able to recognize when they are interacting with AI or encountering AI-generated content.
What content and applications are affected?
Not all content created using AI will have to be labeled in the future. The transparency requirements apply in particular to:
- Certain AI-generated texts on topics of public interest, if there is insufficient human editorial oversight
- AI systems that interact directly with people, such as chatbots
- Emotion Recognition Systems
- certain AI-generated or AI-manipulated content
- so-called deepfakes
Whether labeling is necessary depends on the specific application.
What is considered a deepfake?
The provisions regarding deepfakes are particularly relevant to the creative industry. The AI Act defines a deepfake as AI-generated or manipulated image, audio, or video content that resembles a real person, object, place, or event and could therefore be perceived as genuine.
According to the European Commission’s guidelines, several conditions must be met for this:
- The content resembles a real person, object, place, or event.
- The subject depicted exists or could exist in reality.
- It is an image, audio, or video format.
- The portrayal could be perceived by the audience as genuine or authentic.
Depictions that are obviously unrealistic or purely fictional generally do not fall under this definition.
What are the rules for AI-generated text?
Certain types of content may also be subject to transparency requirements. This applies to content on topics of public interest that has been generated by AI and is no longer reviewed for content or subject to editorial oversight by humans. Simply checking grammar or style is not sufficient in such cases. It is crucial that the content be reviewed and that a human assume responsibility for the published content.
Some questions remain unanswered
There are still some questions of interpretation regarding practical implementation. Guidelines and recommendations are being continuously updated and clarified. At the same time, it has not yet been determined in Austria which authority will be responsible for enforcing the relevant provisions of the AI Act in the future. This question remains open at this time.
Helpful Information Sources
If you’d like to learn more about the new transparency requirements, you can find resources at the following links:
- European Commission Guidelines on the Labeling of AI-Generated Content
- Information pages from the RTR-KI Service Center on the AI Act
- The RTR-KI Service Center’s FAQ section is continuously being expanded
AI remains an important tool for creative work. With the new transparency requirements, the EU aims above all to promote transparency and strengthen trust in AI applications. Those who use AI thoughtfully and keep the new guidelines in mind are well-positioned.