AI in marketing: when must its use be disclosed?

Katri Aarnio
May 27
5 min read

Updated: 3 days ago

Artificial intelligence has entered the world of marketing quickly and with relatively little friction. Images, videos, audio, advertising copy and campaign ideas can now be created in an instant. From a legal perspective, however, this does not mean that the basic rules of marketing have lost their relevance. Rather, AI places familiar questions in a new context. Is the marketing truthful? Does the recipient understand what they are being shown? Who is responsible for the end result?

Which new rules and guidance should marketers pay attention to?

The current discussion is being shaped by three key sets of materials. First, the International Chamber of Commerce, or ICC, has published guidance on the responsible use of AI in marketing. The guidance supplements the ICC’s marketing rules and emphasises that marketing must be lawful, decent, honest and truthful, regardless of the technology used to create it. The ICC guidance is a form of self-regulation rather than directly binding legislation. In practice, however, it may still be relevant when assessing the level of care that can reasonably be expected from a responsible marketer.

Second, Article 50 of the EU AI Act introduces binding transparency obligations relevant to marketing. Many of these obligations will apply from 2 August 2026. They concern, among other things, situations in which a person interacts with an AI system or is exposed to certain content generated or manipulated by AI. Third, the European Commission has published draft guidelines on Article 50, while a Code of Practice on the labelling of AI-generated content is also being prepared. According to the Commission’s draft, the guidelines are intended to provide practical assistance with interpretation. They do not constitute a binding or final interpretation of the AI Act.

There is no automatic obligation to disclose the use of AI

AI can be used in many different ways and at various stages of the marketing process. It is therefore important to recognise that its use does not need to be disclosed automatically in every situation. The ICC guidance states this quite clearly, and the AI Act does not create a general disclosure obligation either. The mere use of generative AI to create advertising materials or marketing communications does not in itself require disclosure. The key question is whether failing to disclose the use of AI could give the recipient a misleading overall impression. Where that is the case, disclosure may be necessary.

Deepfake content is a particular risk area in marketing

Article 50 of the AI Act makes disclosure of AI use a statutory obligation in certain situations. From a marketing perspective, the most relevant example is deepfake content. Under the AI Act, the deployer of an AI system must disclose when image, audio or video content has been generated or manipulated by AI in a way that constitutes a deepfake. According to the AI Act and the Commission’s draft guidelines, a deepfake is content that resembles existing persons, objects, places, entities or events and could falsely appear to be authentic or truthful.

This definition is significant for marketing because deepfakes are not limited to situations involving the imitation of a well-known person. AI-generated content that realistically depicts an existing location, event or product-use context may also fall within the scope of the deepfake rules where the recipient could mistakenly believe it to be genuine. In a marketing context, this could include an AI-generated video showing realistic “customers” using a product in a seemingly authentic setting.

The ICC guidance takes the assessment one step further from a practical perspective. Where AI is used to create or materially alter the image, voice or other likeness of a real and identifiable person for marketing purposes, the person’s permission should generally be obtained and the limits of that permission respected. In practice, marketers must therefore assess two separate questions. They must determine whether they are entitled to use content depicting the person and whether the use of AI must be disclosed to the audience. Article 50 of the AI Act does not directly address consent, as its focus is on transparency. The Commission’s draft guidelines nevertheless point out that deepfake content may also raise issues relating to data protection, intellectual property rights and personality rights.

Disclosure must be assessed from the recipient’s perspective

Another important consideration is the target audience. Both the ICC guidance and the Commission’s draft guidelines on Article 50 emphasise that the clarity of a disclosure must be assessed from the audience’s perspective. In marketing directed at children, older people or other potentially vulnerable groups, the threshold for disclosing the use of AI may in practice be lower. The disclosure must be presented in a way that the particular audience can understand. A purely technical label, a sentence hidden in the terms of use or a vague reference to AI may not be sufficient. According to the Commission’s draft guidelines, the information must be provided clearly and prominently no later than at the time of the first interaction or exposure. It must not be hidden in user instructions or behind a menu structure.

The draft Code of Practice develops this idea further. It suggests that labels for deepfakes and certain AI-generated texts should be easily noticeable, accessible and appropriate for the relevant type of content. The draft also proposes the development of a common EU-wide AI icon and the possibility of a second layer of information explaining in greater detail which elements of the content have been generated or manipulated by AI. From a practical perspective, another interesting proposal is that the label should, where possible, travel with the content when a video or image is shared across different channels.

Responsibility for AI use must be built into marketing processes

For businesses, this means that the use of AI in marketing is not merely a choice of tool for the creative team. The ICC guidance emphasises that the marketer remains responsible for its marketing even where the campaign has been implemented with the assistance of an agency, influencer, platform operator or AI tool. Organisations should update their internal training and instructions so that everyone involved in marketing understands their responsibilities. The same objective is reflected in the AI literacy obligations under the AI Act. Businesses must ensure that personnel using AI have an adequate understanding of its opportunities, limitations and risks.

In practice, a responsible marketer should address three fundamental issues. First, the business should identify where AI is used in the campaign and whether the end result must be labelled under the AI Act. It should then assess whether failing to disclose the use of AI could create a misleading impression, including in situations where the express disclosure obligations under the AI Act do not apply.

Finally, contracts with advertising agencies, influencers and technology providers should support compliance. The marketer should know when AI has been used to create the final output so that it can fulfil its own obligations where necessary.

The use of AI in marketing is not inherently problematic. On the contrary, it can improve quality, accelerate production and create new possibilities for creative work. However, the more authentic AI-generated content looks and sounds, the more important it becomes to consider whether the recipient understands what they are being shown. In this respect, the ICC guidance and Article 50 of the AI Act point in the same direction. Both serve as reminders that trust lies at the very heart of marketing. Where the use of AI undermines that trust, technical efficiency can quickly turn into legal risk and reputational harm.

Businesses should begin preparing early

Final interpretation and enforcement practice are still developing. The Commission’s guidelines on Article 50 and the Code of Practice remain in draft form. The ICC has also stated that it will update its guidance as technology and industry practices evolve. Businesses should nevertheless begin preparing for the obligations now by updating their processes, internal guidance, contracts and approval procedures. From August 2026 onwards, transparency will in many situations become an increasingly concrete and directly applicable legal obligation.