AI disclosure: capabilities and obligations

What is this about?

Article 50 of the EU AI Act regulates transparency obligations for AI-generated content. ALLMEDIA.AI helps you implement them with two tools:

• Notice below posts: A short text such as “AI-generated content.” is automatically appended to the end of the post when publishing. It is not part of the editable text and cannot be accidentally deleted in the editor.
• Labeling of images and videos: Media created in ALLMEDIA receive machine-readable metadata, optionally a credit line in the post and a visible label in the image.

Important: the notice text is switched off initially. You have to enable it once. The media labeling via metadata and credit line, on the other hand, is switched on from the start.

Note: This article explains the features in the product and is not legal advice. Please clarify with your legal advisor for your use case whether and how you need to label content.

Prerequisites

The settings under AI disclosure apply to the whole workspace. They are in the settings area under Content & channels.

Step by step

1. Open the settings. Go to Settings → AI disclosure.
2. Enable the notice text. In the Notice below posts („Hinweis unter Beiträgen“) area, switch on Enable text snippet („Textbaustein aktivieren“).
3. Choose the wording. Under Wording („Wortlaut“), choose a template or Custom text („Eigener Text“, up to 500 characters). The Preview shows what is appended. You choose the text once for the whole workspace.
4. Set up images and videos. In the Images and videos („Bilder und Videos“) area, define whether Write machine-readable metadata („Maschinenlesbare Metadaten schreiben“), Credit line in the post („Nachweiszeile im Beitrag“) and Visible label in the image („Sichtbares Label im Bild“) should be active. For the visible label you choose the Method: Guaranteed overlay (recommended) („Garantiertes Overlay (empfohlen)“) or Prompt instruction only („Nur Prompt-Anweisung“).
5. Deviate per channel. Open a channel in quick edit. In the AI disclosure field, choose Apply workspace default („Arbeitsbereichsvorgabe übernehmen“), Always label („Immer kennzeichnen“) or Don’t label („Nicht kennzeichnen“).
6. Deviate per post. In the post dialog, you set the same choice for a single post in the Details tab. The lowest level that is set wins: post before channel before workspace.

Options at a glance

Option	What it does	Default
Enable text snippet	Appends the notice to the end of the post when publishing, after the image credits	off
Wording	“This post was created with the support of artificial intelligence and reviewed editorially.” (recommended), “Parts of this post were generated using artificial intelligence.”, “AI-generated content.” or custom text	first template
Write machine-readable metadata	Writes the XMP/IPTC label “trainedAlgorithmicMedia” into the image file	on
Credit line in the post	Adds “Image: AI-generated (model)” („Bild: KI-generiert (Modell)“) to the image credit	on
Visible label in the image	Labels generated images directly in the image	off

The notice text appears in the channel’s language. In the post preview you see it as a locked line.

Platform-specific labeling

Some platforms offer their own AI labels, for example TikTok, Instagram or YouTube. The labeling in ALLMEDIA does not replace them. Check beforehand whether the platform additionally requires its own label for your content.

Limits

• The notice text does not depend on whether a post was actually created with AI. If it is active, it is appended to all posts in the respective scope. For purely manually written posts, you therefore set Don’t label in the post.
• You can only define the wording for the whole workspace. Channel and post only switch the notice on or off.
• Image models do not reliably implement the visible label via Prompt instruction only. The overlay, the metadata and the credit line are more reliable.
• Media labeling applies to media created in ALLMEDIA. In the Gallery they are marked AI-generated.

Your responsibility in practice

• Check whether and how you need to label content in your case, and set up AI disclosure accordingly.
• Also use the platforms’ AI labels where they are offered.
• Check content for accuracy before publishing. Labeling does not replace fact-checking or clarifying image rights.

Frequently asked questions

Why don’t I see the notice in the editor text?

It is not part of the content but is only appended when publishing, just like image credits. In the preview it appears as a locked line.

No notice appears below my post. Why?

First check whether Enable text snippet is switched on. Then check channel and post: if Don’t label is set there, it overrides the workspace default.

Related articles

• Data privacy and GDPR in ALLMEDIA.AI
• What you can publish per channel
• AI Studio: AI images, videos and carousels
• Channel canvas and approval workflow