Not every AI assisted sentence needs a banner. Article 50 draws lines around specific kinds of AI generated content. Get those lines right and you avoid both under labelling (a compliance gap) and over labelling (slapping a warning on everything, which trains readers to ignore it).
Two different labelling duties
| Duty | Who | How it shows |
|---|---|---|
| Mark synthetic output (50(2)) | Provider of the generative system | Machine readable, embedded where feasible |
| Disclose deepfakes (50(4)) | Deployer publishing the content | Visible to the viewer, at first exposure |
| Flag public interest text | Deployer publishing the text | Visible disclosure that the text is AI generated |
The public interest text rule
AI generated text published to inform the public on matters of public interest must be disclosed as AI generated. A newsroom auto generating a report on an election is the model case. Purely internal drafts, or text a human editor has reviewed and taken responsibility for, are treated differently. The trigger is publication to inform the public on public interest topics.
Visible label vs machine readable mark
These are not the same thing and you may owe both. The machine readable mark (50(2)) is invisible metadata a provider embeds so downstream systems can detect the content is synthetic. The visible disclosure (50(4) and the text rule) is what a human reader sees. As a publisher, your job is the visible side.
Avoid over labelling
Labelling everything, including content a human wrote and owns, dilutes the signal and annoys readers. Reserve the disclosure for what the law actually targets: realistic synthetic media, deepfakes, and public interest AI text.
Common questions
I use AI to draft blog posts a human then edits. Do I have to label them?
The strong trigger is AI generated text published to inform the public on public interest matters. A human edited marketing post that a person takes responsibility for is a different case. When the topic is public interest and the text is machine produced, disclose it.
Is a footer note enough for an AI image?
For a realistic synthetic image, the disclosure must be clear and given at first exposure. A visible label on or near the image is stronger than a note buried in the footer.
Label AI content the right way
The widget's content label templates cover synthetic media and AI text, localized and shown at first exposure, so you disclose what the law targets and nothing more.
See pricing