Artificial intelligence can create advertising copy, images, logos, and videos in seconds. But just because something is generated quickly doesn’t automatically mean it’s legally compliant. Companies must pay particular attention to copyright, trademark, competition, and data protection laws. In addition, new transparency requirements under the EU AI Act have been in effect since August 2, 2026.

AI advertising is permitted, but it is not without risks

The use of generative AI in marketing is generally permitted. However, the company that publishes the content remains legally responsible. It should therefore not rely on an AI tool to deliver only unobjectionable results. AI-generated product names, logos, slogans, or images may be too similar to existing trademarks, copyrighted works, or distinctive product designs.

Carefully review copyright and trademark rights

AI systems are trained using large amounts of data and can generate content that resembles existing photographs, texts, designs, or trademarks. If a protected work is copied or imitated too closely, this may result in claims for injunctive relief, damages, and warning letter costs. Even a merely similar name or design can be problematic if it creates a risk of confusion. The extent of this risk is illustrated by the decision of the Munich I Regional Court regarding AI training using protected music.

Conversely, not every result generated entirely by AI automatically enjoys copyright protection. As a general rule, a work must be a personal intellectual creation—that is, it must involve a sufficient human creative contribution—to be protected by copyright. If this is lacking, it can be difficult to prohibit third parties from using or imitating an AI-generated design. Our article on AI-generated logos illustrates what this means in practice.

Especially when it comes to key campaign visuals, logos, or advertising graphics, companies should therefore clarify early on what rights they actually hold. In doing so, they must take into account both the terms of use of the respective AI provider and any potential third-party rights. If a design is to be used exclusively over the long term, it may be advisable to have the AI-generated results further refined by a designer and reviewed for legal compliance.

Misleading advertising remains prohibited

AI-generated advertising must also be truthful and must not be misleading. Fictitious product features, fabricated customer testimonials, or realistic-looking depictions must not deceive consumers. For example, anyone who features a nonexistent person as a supposed expert, customer, or testimonial must carefully assess the impression the advertisement conveys.

In addition, imitating specially designed products or packaging may be prohibited under competition law. This applies even if no trademark or copyright is infringed.

When must AI-generated content be labeled?

The transparency requirements set forth in Article 50 of the EU AI Act have been in effect since August 2, 2026. However, there is no general requirement to visibly label every AI-generated text or image as “AI-generated.” The obligation to label content that is deceptively realistic falls on the publisher (Article 50(4)). There is no transition period for this requirement.

In particular, companies must disclose when AI-generated or significantly manipulated image, audio, or video content could be mistaken for real people, objects, places, or events and thus constitutes a deepfake. Photorealistic advertising content may also fall under this requirement.

AI-generated texts are subject to a labeling requirement, particularly when they inform the public about matters of public interest and have not undergone sufficient human or editorial review. This is distinct from the provider’s obligation under Article 50, paragraph 2: Synthetic content must be marked in a machine-readable format at the time of creation. For systems placed on the market before August 2, 2026, the AI Omnibus Directive provides for a transition period until December 2, 2026. However, those who use third-party AI tools cannot invoke this provision, as their own labeling obligation exists regardless of whether the system they use already technically labels the content. Machine-readable labeling and human-perceivable labeling are two different things; one does not replace the other. We have traced the evolution of the discussion on labeling in our article on AI labeling.

Don’t forget about privacy

Anyone who analyzes personal data, customer profiles, or usage patterns using AI needs a legal basis under data protection law. Confidential customer, employee, or company data should not be entered into freely accessible AI tools without first being reviewed. Depending on the use case, privacy notices, data processing agreements, and technical safeguards may also be required.

Legally compliant AI advertising requires oversight

AI can accelerate marketing processes, but it does not replace legal and editorial review. Companies should establish binding internal policies, document the tools they use, review results for third-party rights and misleading statements, and clarify any potential disclosure requirements before publication.

In our podcast episode, “Using AI Images Legally: But How?,” we discuss how to use AI images in a legally compliant manner in practice.