AI commercials and the legal implications of synthetic reality

Commercials have always worked with imagination. AI changes the scale and quality of this imagination. It allows advertisers to create marketing materials that look realistic and technically possible, while being fully synthetic. Advertisers understand this value and widely adopt AI for their campaigns, often without paying enough attention to the legal implications of such use.

However, this creates several legal risks.

First, if the commercial creates an impression that the product can achieve something, the advertiser should be able to prove it. Otherwise, such advertising may be considered misleading and lead to liability.

A useful example is the recent BBC investigation into AI fitness adverts[1]. It described fitness apps using AI-generated instructors. A fictional demonstration may be understood as proof that the advertised result is achievable. In reality, consumers may not be able to achieve the unrealistic results shown, even if the ad creates the opposite impression.

This is where the second legal implication arises. If AI content can be perceived as a factual representation of reality, the question is not only whether the claim is true, but also whether the synthetic nature of the content is sufficiently clear.

The UK Advertising Standards Authority ruling concerning a robot-dog toy ad[2] shows why clarity about the nature of the content becomes a legal risk. The ad promoted the toy through scenes of puppies behaving like real dogs. The scenes appeared to be real and/or AI-generated puppies and created the impression that the toy itself was realistic and lifelike. In reality, the product did not behave in that way.

A clear marking or disclaimer could have reduced the misleading impression, for example: “The video footage is AI-generated. The real product may differ in appearance and functionality from what is shown.” However, no such clarification was provided. As a result, consumers could reasonably understand the synthetic footage as a factual demonstration of the toy’s characteristics.

The same synthetic freedom creates a third risk. Even where the commercial does not misrepresent the product, AI may still generate content that the advertiser does not recognise as someone else’s intellectual property. The risk is not limited to obvious cases where Mickey Mouse is used[3]. AI may generate a character that looks neutral, but is in fact close to someone else’s protected work. The company often does not know where the AI system took the idea and visual expression from, making the risk difficult to identify in advance and hard to exclude.

This is especially important in unfamiliar markets. The Ultraman AI case in China[4] shows the problem: generative AI can reproduce or approximate protected fictional characters. The final result may look like a new imaginary character, while still being legally too close to an existing copyrighted work.