AI images, and the new EU AI Act

Under the EU AI Act, transparency rules for AI-generated/manipulated content began applying on 2 August 2026.

What are your plans regarding this? Are you finally considering creating a separate category specifically for AI-generated content?

I’m just asking as a regular user and would be interested to know what you’re planning.

Thanks for asking. A separate category isn’t as straightforward as it might seem: several of our theme stores are consumed by external services, for example the KDE Plasma desktop, so any structural change on our side has to work for those integrations as well.

Our current approach is to ask users to label their content with one of two lowercase tags:

  • ai-generated – for content that was created entirely or mostly by AI
  • ai-assisted – for content where AI was used as a tool, but the work is substantially the uploader’s own

Please use exactly these tags (all lowercase, with a hyphen) so that filtering works consistently.

This does two things: it allows KDE to filter such content out on the Plasma desktop if they choose to, and it lets us offer corresponding filters in the store itself.

As with all content, accurate labelling is the uploader’s responsibility – we can’t reliably detect AI-generated content on our side.

That’s where things stand today – no guarantees implied. If the legal situation develops further, we’ll adapt accordingly.

Thanks for your quick response, dschinn-de!