Germany: New EU AI Rules Require Labelling of Certain AI-Generated Images from August 2026

From 2 August 2026, AI-generated images published beyond purely private use may need to be labelled under certain circumstances pursuant to Article 50 of the EU AI Act (Regulation (EU) 2024/1689).

Who is subject to the obligation?

Under the AI Act, the labelling obligation applies to any “operator” of an AI system capable of generating images. Pursuant to Article 3(4) of the Regulation, an operator is any natural or legal person using an AI system under their own responsibility, unless the use is purely personal and non-professional.

In practice, this means that companies and individuals using AI-generated images for professional purposes may fall within the scope of the obligation. This includes the use of AI-generated content on corporate websites, marketing materials, and professional social media platforms such as LinkedIn.

Which images must be labelled?

Article 50(4) of the AI Act requires the labelling of certain “deepfakes”. The definition is considerably broader than the term’s common everyday usage.

For the purposes of the Regulation, a deepfake includes AI-generated or substantially AI-manipulated image, video, or audio content that resembles real persons, places, objects, or events and could reasonably be mistaken for authentic content.

Importantly, the decisive factor is the likely perception of the audience rather than the creator’s intention.

Labelling obligation: practical examples

Not every AI-generated image will qualify as a deepfake. The key question is whether an average viewer could reasonably believe the image to be genuine.

Examples likely requiring labelling:

• Realistic image of a fictional person or object (e.g. a woman working at a desk, a tree in a meadow, fish underwater)

• AI-manipulated photograph of a real person (e.g. a public figure depicted in an altered context)

• AI-generated image of a real building shown in a fictitious condition

• AI-generated image of a fictional event or location presented realistically

Examples unlikely to require labelling:

• Clearly stylised logos or abstract AI-generated graphics

• Product mock-ups in comic or illustration style

• Obviously fantastical scenes (e.g. dragons, fictional planets, witches flying on broomsticks)

• Caricatures or satirical images that are clearly exaggerated or unrealistic

Practical rule of thumb

A useful practical test is whether an impartial observer could mistake the image for a real photograph or authentic depiction.

If the answer is yes, the content is likely to qualify as a deepfake under the AI Act and should be labelled accordingly. If the image is obviously artificial, fantastical, or heavily stylised, the obligation may not apply.

As there is currently no established case law interpreting these provisions, caution is advisable in grey areas such as AI-enhanced retouching, AI-generated backgrounds, or AI-sharpened photography. In cases of uncertainty, labelling is generally the safer approach.

What should the labelling look like?

The labelling must be clear and easily recognisable to viewers. In practice, this may include a visible caption, watermark, overlay, or other prominent disclosure indicating that the content was generated or materially altered using AI.

Potential consequences of non-compliance

The AI Act provides for administrative fines of up to EUR 15 million or 3% of global annual turnover, whichever is higher, for certain violations of Article 50.

In addition, failure to label qualifying AI-generated content may also give rise to claims under German unfair competition law (UWG), particularly where the omission could be regarded as misleading commercial conduct.

By MELCHERS, Germany, a Transatlantic Law International Affiliated Firm.

For further information or for any assistance please contact germany@transatlanticlaw.com

Disclaimer: Transatlantic Law International Limited is a UK registered limited liability company providing international business and legal solutions through its own resources and the expertise of over 105 affiliated independent law firms in over 95 countries worldwide. This article is for background information only and provided in the context of the applicable law when published and does not constitute legal advice and cannot be relied on as such for any matter. Legal advice may be provided subject to the retention of Transatlantic Law International Limited’s services and its governing terms and conditions of service. Transatlantic Law International Limited, based at 84 Brook Street, London W1K 5EH, United Kingdom, is registered with Companies House, Reg Nr. 361484, with its registered address at 83 Cambridge Street, London SW1V 4PS, United Kingdom.