A Guide to Using the New EU Icons for AI Content | Part 1

A Guide to Using the New EU Icons for AI Content | Part 1

Learn how to use the new EU icons for AI content. Discover the differences between basic, fully generated, and partially modified icons for AI compliance.

AI Video War: Sora 2 is the 'World Builder', But Gen-3 and Seedance Have Their Own Superpowers Reading A Guide to Using the New EU Icons for AI Content | Part 1 3 minutes

With the introduction of the AI Act's transparency rules, content creators and publishers face a new challenge: how to clearly and effectively disclose artificial manipulation. To simplify this process, the European Commission has introduced a standardised set of EU icons for AI content. By using these freely available visual markers, deployers of generative AI can ensure their audience remains informed while supporting public trust. Here is a comprehensive guide on how to understand and implement these new icons.

Introducing the EU Icons for AI Content

The European Union has developed a specific set of icons that function as an integral part of the Code of Practice on marking and labelling of AI-generated content. These EU icons for AI content were designed following rigorous user testing, which proved that visual markers—especially when accompanied by clear text labels—significantly improve audience recognition.

The icons are available in four variations to suit different design needs:

  • Solid black

  • Solid white

  • Black with 50% transparency

  • White with 50% transparency

Creators can download these files in both SVG and PNG formats, making them easy to integrate into any digital medium.

The Three Main Icon Categories

To provide accurate context, the EU icons for AI content are divided into three distinct usage categories.

1. The Basic Icon

This icon is the most versatile. It should be used when AI was involved in the creation of deepfake content (image, audio, or video) or published text. It is also the correct choice when a custom text label or an interactive second layer of information is implemented. For instance, a deepfake video might feature the text "voices generated with" followed by the basic icon.

2. Fully AI-Generated

When a piece of content is entirely fabricated by an AI system with no human-created elements or editorial control (aside from the initial prompting), this icon must be used.

  • Examples: Fully AI-composed music, fully AI-generated news summaries, or deepfake videos depicting fictional events.

3. Partially AI-Modified

This icon is required when pre-existing, human-made content is altered using AI.

  • Examples: Swapping a face in an authentic photograph using AI, or using AI tools to furnish an authentic photograph of an empty apartment.

Best Practices for Displaying the Icons

While using these specific EU icons for AI content is optional, the labelling requirements under Article 50 of the AI Act are mandatory. If you choose to use these icons, they must be displayed correctly:

  • They must be clearly perceivable at the first moment of exposure.

  • They should be directly embedded into the deepfake or published text so they remain visible if the content is downloaded or reshared.

  • They must not be obscured by intervening overlay elements.

Link to download the official icons from the EU AI Office.

By standardising how we mark artificial media, these EU icons for AI content make compliance straightforward. In the final part of our series, we will delve into the exceptions to these rules and the importance of accessibility.

Leave a comment

All comments are moderated before being published.

This site is protected by hCaptcha and the hCaptcha Privacy Policy and Terms of Service apply.