MarTech & AI

A Neural Network Made the Banner: Who Owns the Rights to AI Creative

48% of companies use AI in advertising, but a third have no rules for its use. Together with a legal counsel at Bitrix24, we break down who owns the rights to generated banners and slogans, and which creatives will lead to lawsuits.

4 min read

Кому принадлежат права на рекламные материалы, созданные нейросетью
Кому принадлежат права на рекламные материалы, созданные нейросетьюAI-generated

Nearly half of Russian companies (48%) are already using artificial intelligence and machine learning in advertising campaigns. Generative AI for marketing is no longer an experiment: neural networks create banners, videos, product images, texts, and presentations, reducing content production time and costs.

The speed of technology adoption is noticeably outpacing the development of internal policies. More than 80% of Russian companies allow the use of generative AI in work processes, yet almost a third of organizations (32.2%) impose no requirements for the safe use of such tools — even though 95% of businesses acknowledge the associated risks. Companies automate the creation of advertising materials, but far from always do they check who owns the rights to the generated content, whether the license permits commercial use, and whether the creative infringes on third-party rights.

Danil Shimko, legal counsel at Bitrix24, explains how to build a safe workflow with generative AI.

Law and Practice

It should not be assumed that copyright and exclusive rights to an advertising creative automatically belong to whoever wrote the prompt for the neural network. Russian legislation does not directly regulate authorship of content created with AI. Under the Civil Code of the Russian Federation, the author of a work can only be the citizen whose creative labor produced it. If a neural network fully generated an image or text from a prompt without any creative human contribution, such a result may not be protected by copyright.

It is a different matter if a person used AI as an additional tool: reworking the result, changing the composition, adding their own elements. In that case, one can speak of a creative contribution, and protection may extend to the final material in the relevant part. The mere fact of writing a prompt does not guarantee that the user will be recognized as the author of the generated result.

Example: a marketer generates a banner with a neural network, after which a designer completely reworks the composition, replaces elements, adds brand graphics, and adapts the image to different formats. Here, AI is merely a tool, and the final result includes human creative input. A different situation: an image obtained from a simple prompt like "modern office, photorealism," unchanged, goes straight into a campaign. Copyright to such material most likely does not belong to the user who wrote the prompt: there is no creative labor.

For business, the key is acquiring exclusive rights to the created material or license agreement terms permitting commercial use of the results. Before launching a campaign, it is worth checking whether the chosen service allows commercial use of generations and whether it transfers exclusive rights. Both the license terms and the tariff plan under which the material was created matter.

AI "Came Up With" a Slogan: Can It Be Protected From Copying

Risks differ for different types of advertising content. With slogans, an additional question arises — whether they can be given legal protection at all. In some cases, a slogan can be registered as a trademark, but only if it has distinctive character and does not coincide with existing marks.

If a neural network suggests the slogan "The Future Has Already Arrived" or "Technology That Works for You," protection is hardly to be expected: the phrasings are too generic. The situation is different with complex, original marks — those can be an attempt to register as trademarks. Before launching a campaign, it is worth checking not only the originality of the phrasing but also the existence of similar trademarks. If a slogan is a long-term brand element, the legal review is best conducted before the campaign goes to market.

Deepfake in Advertising: Where the Line Between Creative and Lawbreaking Lies

A separate category of disputes involves materials using deepfakes. Replacing an actor's face with the image of a famous person, creating a digital copy of a celebrity, or reproducing the characteristic features of a popular character can lead to claims from rights holders or the person themselves. Risk arises even if the ad does not use an original photograph but merely creates a similar image with a neural network.

For example, the Smeshariki franchise has images of virtually all its main characters registered as trademarks; Soyuzmultfilm takes a similar approach. This allows the characters to be protected under several legal mechanisms at once. Lawyers recommend avoiding even recognizable references, not just direct copying.

Similar risks apply to the use of digital likenesses of real people. In 2026, an illegal bookmaker distributed an AI-generated video featuring footballer Bruno Fernandes, creating the impression that the athlete was the face of the brand. No partnership existed, and the case became one of the most discussed examples of unlawful use of a person's digital likeness for commercial purposes.

Which Images Cannot Be Used Even After Generation

Even a fully fictional character created by a neural network is no guarantee of safe use. Generative models reproduce recognizable elements of existing works, and a company risks facing claims after the campaign has already launched. At risk are popular characters, logos, and brand images: trying to generate an "almost Batman," a "hero in the style of Marvel," or a character with the features of well-known franchises does not eliminate the risk of infringing intellectual property rights.

In practice, the prompts look harmless: "create a superhero in the style of Batman, but in different colors," "draw a wizard who looks like Harry Potter," "make a yellow cartoon character like a Minion." Formally, the neural network generates a new image, but if the audience easily recognizes a specific character, the rights holder may have grounds for claims.

In June 2025, Disney and Universal Pictures filed a lawsuit against Midjourney, claiming that the service allows the creation of images practically identical to famous characters — from Darth Vader and Shrek to the Minions and Marvel heroes. Even an image created from scratch that reproduces a recognizable character without permission infringes the rights of rights holders.

A similar approach applies to characters of folklore. The fairy-tale image of Baba Yaga can be freely interpreted — for example, a restaurant may use its own original illustration of the character in an autumn campaign. But if a neural network creates an image clearly replicating the appearance of the character from a specific cartoon or a famous artist's illustration, such use may infringe the rights holder's rights.

Legal Audit of AI Creatives: What to Check Before Launch

It is no longer enough for companies to assess only the creative concept or compliance with the brand book. Generative AI requires a separate legal review: it helps identify in advance intellectual property infringement, non-compliance with license terms, use of third-party images and copyrighted objects, and assess the content's compliance with legal requirements.

Source: Cossa; Source: cnews.ru

Share this article:

More from Екатерина Ларина

Related articles

« Previous articleNext article »