An AI chatbot can streamline customer inquiries and reduce the workload on employees. However, if it disseminates incorrect information, the company is liable: The Higher Regional Court of Hamm ruled (OLG Hamm, judgment of May 12, 2026, Case No.: 4 UKl 3/25) that operators cannot claim that “the AI” made the mistake. Anyone who uses a chatbot on their own website must take responsibility for its statements.

What was the case before the Higher Regional Court of Hamm about?

The Higher Regional Court (OLG) of Hamm had to rule on a case that is significant for many companies with online services. A cosmetic surgery practice used an AI chatbot on its website that allowed customers to ask questions or get in touch. Among other things, the chatbot referred to the managing directors as “specialists in plastic and aesthetic surgery” or used similar specialist titles.

The problem: While the title “Specialist in Plastic and Aesthetic Surgery” does exist, the managing directors had not completed the required training. The titles “Specialist in Aesthetic Medicine” and “Specialist in Aesthetic Treatments” do not exist as medical specialties at all. For patients, however, such a claim can be very important. Those seeking information about medical or aesthetic treatments often pay particular attention to qualifications, specialist titles, and professional experience. A false specialist title can therefore give the impression that a specific medical qualification exists, even though this is not actually the case.

The North Rhine-Westphalia Consumer Protection Agency took legal action against these statements. The Higher Regional Court of Hamm upheld the complaint and made it clear that the chatbot’s false statements could be attributed to the company.

Why is this decision so important?

The ruling is particularly interesting because it isn’t just about a single false statement on a website. It raises a fundamental question: Who is liable when an AI chatbot provides false information?

Many companies now use chatbots to automatically answer customer questions. This applies not only to medical practices and cosmetic clinics, but also to online stores, service providers, insurance companies, law firms, real estate agents, travel agencies, gyms, and many other industries. Especially when a chatbot provides information on services, prices, qualifications, contract terms, or legal issues, errors can have significant consequences. We address fundamental questions of liability regarding the use of AI in our article Liability Issues in the Use of AI in Business.

The Higher Regional Court of Hamm has made it clear: The chatbot is not an independent third party. It is a tool of the company. Whoever uses this tool must also control it. Companies cannot, therefore, use the fact that a false statement was generated solely by AI as a defense.

What does this mean for companies that use AI chatbots?

For businesses, this decision means one thing above all else: AI chatbots must not be used unsupervised as digital information providers if they make statements relevant to customers. Anyone using a chatbot on their own website should ensure that the responses are both legally and factually accurate.

Statements regarding professional qualifications, certifications, approvals, prices, warranties, delivery times, medical services, success rates, or the specific benefits of an offer are particularly risky. Such information can be decisive for consumers’ business decisions. If it is false or misleading, this can give rise to claims under competition law.

This is true even if the incorrect statement was not written by a human. The key point is that the chatbot is used on the company’s website and communicates there on behalf of the company. From the customers’ perspective, the response appears to be part of the company’s official presence.

Why can an incorrect chatbot response be anti-competitive?

One of the key objectives of competition law is to protect consumers from misleading commercial practices. A commercial practice can be problematic if it is likely to influence customers to make a decision they might not otherwise have made.

For example, if a chatbot claims to have a specific qualification that does not actually exist, this can boost trust in the company and influence a decision to seek treatment, enter into a contract, or make contact. This is precisely where the legal risk lies.

The Higher Regional Court of Hamm therefore did not view the chatbot’s statements as a mere technical glitch, but rather as statements made in the course of business. The chatbot was not communicating privately or incidentally, but rather as part of the company’s online platform. As a result, its responses were attributed to the company.

“It was the AI’s fault” isn’t enough

For those without a legal background, the key message of the ruling can be summarized simply: Anyone who uses AI in their own company remains liable.

A company cannot hide behind technology. A regular employee, an external service provider, or an advertising agency can also provide false information. Nevertheless, in many cases, the company remains responsible for its public image. The Higher Regional Court of Hamm takes a similar view regarding AI chatbots: The technology does not operate independently of the company, but is used by it to facilitate customer communication.

AI systems, in particular, carry the risk of so-called “hallucinations.” This refers to situations where an AI generates answers that appear convincing but are factually incorrect. For consumers, such errors are often difficult to detect because the answers appear linguistically confident and professional. This makes it all the more important for companies to establish clear boundaries and control mechanisms.

What Companies Should Be Reviewing Now

Companies should first analyze which tasks their chatbot handles. Does it only answer general service questions, or does it also provide information on products, prices, qualifications, legal terms, or medical services? The more sensitive the content, the stricter the controls should be.

It makes sense to establish clear limits on what the chatbot can say. A chatbot should not be allowed to make up qualifications, titles, guarantees, or legally binding statements. When in doubt, it should refer the user to a human representative. In addition, companies should regularly test, document, and adjust the chatbot’s responses.

A clearly visible notice indicating that the service is an AI chatbot can also be helpful. However, it does not replace content moderation. Transparency alone does not protect against liability for false or misleading statements.

AI lightens the load—but it does not relieve us of responsibility

The ruling by the Higher Regional Court of Hamm shows that using AI in customer interactions is not a sure thing. Chatbots can help lighten the workload, but they also pose legal risks if they make relevant statements without proper verification.

Companies should therefore establish clear guidelines, conduct regular reviews, and set rules for sensitive topics such as advertising, pricing, or service promises. This is because statements made by the chatbot can be attributed to the company.

The key message: AI can provide support, but it does not relieve companies of their responsibility.

In our podcast episode “AI Labeling: The Second Draft,” we discuss the labeling requirements that will apply to the use of AI starting in August 2026.