Avoiding AI-washing: Legally compliant advertising with artificial intelligence
Advertising with the term "AI" may be misleading. In this article, we will show when "AI-washing" occurs and how it can be avoided.
Advertising for artificial intelligence (AI) is widespread. More and more enterprises are emphasising the use of AI in their enterprise or the integration of AI into their products in their advertising, for example by describing their products as "AI-driven" or "AI-supported". This allows them to present themselves or their products as particularly innovative. AI is even being used in everyday objects, such as washing machines or televisions – if the advertising claims are to be believed.
Such advertising is not objectionable if the information advertised reflects the facts. In many cases, however, there is a suspicion that "AI-washing" is taking place. AI-washing (a term borrowed from terms such as "greenwashing") refers to a misleading marketing practice in which enterprises claim that AI is used in their products or in the provision of their services even though no AI is used at all or this is only true to a very limited extent. The AI hype is being exploited.
Competition law risks of AI-washing
AI-washing can constitute misleading advertising in accordance with section 5 (1) and (2) no. 1 German Unfair Competition Act (UWG) and therefore be anti-competitive. This is the case if the advertised product does not even have the advertised AI functions. It is also anti-competitive if the advertising creates expectations that the product does not fulfil. Enterprises that make false or even misleading promises in this way are not only deceiving their customers – they are unfairly gaining a competitive advantage over their competitors.
AI-washing has already come under the focus of consumer organisations and authorities. In the US, for example, two investment advisors have been targeted by the US Securities and Exchange Commission (SEC) for making false and misleading statements about the alleged use of AI. In the UK, the Advertising Standards Authority (ASA) has criticised the advertising for an app that can be used to edit images – allegedly using AI – for being misleading. The advertisement exaggerates the app's performance.
It is probably only a matter of time before consumer organisations in Germany start to take action against "AI-washing" on a larger scale, as can be observed in the case of greenwashing.
AI-washing and the term "artificial intelligence"
There is no general answer as to when advertising of the use of AI is misleading and therefore constitutes AI-washing under competition law. The first reason for this is that there is no general common understanding of the term AI. There is no single "AI", but rather different types of AI, such as generative AI. When trying to answer the question of whether an AI advertisement is misleading, it is not just a matter of a technical understanding of the term. Instead, the decisive factor is the impression made on the relevant public based on the wording of the statement in the advertisement and the overall context in which the advertisement is placed. Ultimately therefore, it depends on the circumstances of the individual case.
One way of understanding what is meant by AI may be the definition of AI systems in the AI Act. It is recognised that public perception can be influenced by legal provisions, for example if the content of definitions is defined by law ("purified public perception"). According to Article 3 no. 1 AI Act, an "AI system" is:
"a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments."
AI systems and AI in general are characterised by the fact that they are "machine-based", "autonomous", "adaptive" and "capable of influencing environments".
Machine-based means the action of a machine. For example, selecting a washing machine programme or selecting how much washing powder to use is machine-based if this is done by or with the help of a machine and not exclusively by a person.
Autonomous means that the system operates independently of humans. Once it has been put into operation by a human, the system makes decisions independently without any further human involvement from this point onwards. It decides independently how to proceed on the basis of the data available to it. In the example of the washing machine, this would be the independent decision on a specific washing machine programme or the independent decision on how much washing powder to use.
A system is adaptive if it learns as it is used. This characteristic distinguishes AI significantly from simple software and conventional programming approaches, which are based on rules defined by natural persons. For example, a washing machine that "learns" based on user behaviour or other information and develops accordingly is "adaptive" in this sense. If, by contrast, the washing machine just has sensors installed in it to determine the type of fabric or the weight of the laundry, based on which it starts a certain programme or determines how much washing powder to use, this does not meet the definition. In this case, a human being specified what was to be selected when a certain condition determined by a sensor was met.
The last attribute, being capable of influencing environments, represents being independently able to actually execute the decisions previously made by the machine. This could be, for example, automatically starting the previously selected washing machine programme and actually starting the machine independently.
If at least one of these features is not present, it is reasonable to assume that the customer's expectations of the advertised "AI product" will not be satisfied and that this is therefore a case of AI-washing with relevance to competition.
However, it should also be noted that even if a product advertised with some reference to AI can be classified as an AI system within the meaning of the AI Act, consumers or other customer groups may still be misled by advertising. This is particularly the case if potential customers are given the impression that the product contains certain AI functionalities when this is not the case or if the product has only a marginal AI component. An example of this is when a software program for detecting skin cancer that is advertised as "AI-enabled" is indeed based on a trained neural network, but – contrary to the expectations of the target customer groups – does not undergo continuous improvement through new data. This creates the impression that the entire diagnostic process is highly automated and capable of learning, though this is not actually the case.
If you want to avoid AI-washing, you need to communicate clearly and transparently
Misleading advertising for AI is associated with the risk that consumer associations or competitors may file a claim for injunctive relief, compensation or confiscation of proceeds. However, AI-washing can also be a criminal offence and result in investigations by the public prosecutor's office. This may be the case, for example, if there is a suggestion that customers have been deceived and induced to make a purchase, i.e. if there are indications of fraud. In any case, it can lead to disputes with customers and damage to reputation. Enterprises that advertise an "AI product" or "AI service" should therefore bear in mind that this can lead to accusations that they are only using AI as an advertising ploy. To avoid being accused of this, a few things should be taken into account:
It goes without saying that advertising should only use terms such as "artificial intelligence", "AI" and "AI-supported" if AI is actually used. As the term AI is vague and can therefore lead to exaggerated expectations and misunderstandings, it is advisable to provide transparent and concrete information about the use of AI in the product. It is not advisable to use meaningless advertising promises with the buzzword "AI" such as "Our product is powered by AI!". Instead, enterprises should emphasise the concrete added value that AI offers compared to when "simple" software is used in the product. It is recommended to describe specifically what the AI does and to communicate where the limits of the AI lie. When doing so, enterprises must take into account whether their advertising is aimed at consumers or professionals. In addition, it should be possible to prove that the advertising claims are accurate, i.e. that AI is actually used in the manner described. Enterprises should therefore maintain appropriate documentation that can be referred to if an accusation of AI-washing is made.
In this context, enterprises should also check whether they comply with the transparency and disclosure obligations applicable to AI. Some transparency and information obligations will apply from 2 August 2026 in accordance with the AI Act. This regulation governs various circumstances in which reference must be made to the use of AI. A breach of the regulation can be penalised with fines of up to EUR 15,000,000 among other sanctions. In addition, a general obligation to provide information on the use or deployment of AI can also be derived from section 5a (1) German Unfair Competition Act (UWG). According to section 5a (1) German Unfair Competition Act (UWG), anyone who omits material information is deemed to have acted unfairly. Information is material if it is necessary for the consumer to make an informed business decision. The use of AI can be seen as material information, for example when user data is used for AI training purposes.
Careful examination of advertising with AI is unavoidable
Advertising with AI must be checked carefully to avoid legal risks such as "AI-washing". Clear, transparent communication about the use of AI is crucial to avoid giving false expectations, prevent formal warnings and not lose the trust of customers. Only by presenting AI functions in a precise and comprehensible manner can enterprises prevent competition law disputes and damage to their reputation.
