Background altered using AI, lighting adjusted, distracting detail removed: in marketing, this is done in no time. The transparency rules set out in the AI Regulation have been in force since 2 August 2026. However, the regulation does not require every AI-edited image to be visibly labelled.

What matters when it comes to photos

A disclosure is particularly necessary when an AI-edited photo is a deepfake: it resembles a real person, object, place or event and could be mistaken for the real thing. In such cases, it must be disclosed that the image has been artificially generated or altered.

A realistic-looking background inserted using AI may therefore be relevant – for example, if it gives the impression that a product was photographed in a location where it has never actually been. Whether this constitutes a deepfake in a specific case depends on the overall image and its effect.

Minor alterations do not automatically require labelling

A colour or lighting correction, cropping or the removal of a distracting detail does not automatically make a photograph a deepfake requiring labelling.

And what about text and music?

There is a duty to disclose AI-generated text if it informs the public about matters of public interest. This requirement does not apply if the text has been checked by a human or subject to editorial control, and a natural or legal person bears editorial responsibility. AI-generated music does not, as a general rule, need to be visibly labelled. An AI voice that imitates a real person so convincingly, however, may fall under the deepfake rule. In the case of AI chatbots, it must generally be clearly recognisable that one is speaking to an AI system. This does not apply if it is already obvious in the specific context.

Would you like to know whether and how the use of AI in your advertising must be disclosed, and what other framework conditions must be observed when using AI-generated marketing materials? Our experts in AI and IP/IT at KWR will be happy to advise you.