Munich Court Holds Google Liable for False AI Overviews: What the Preliminary Injunction Means for Insurers' Chatbots

A German court has said that a technology company is responsible for what its artificial intelligence says, and insurance professionals are paying attention. On May 28, 2026, the Regional Court of Munich issued a temporary injunction against Google over its AI Overviews feature, according to an Insurance Business analysis headed "What the Munich Google ruling means for insurers." This article summarises that analysis. An earlier version of this article said that Google confirmed on June 12 that it would appeal. The report we reviewed says Google stated on June 11 that it is reviewing the decision, which is not yet final, and it does not describe a specific appeal, so the article now reflects that. The ruling is a preliminary measure and the case may develop further, so readers should check for current news.

What happened? The court issued a temporary injunction barring Google from repeating false statements about two Munich-based publishers. AI Overviews, Google's feature that shows AI-written summaries at the top of search results, had linked the publishers to scams and dubious business practices. Those claims appeared in none of the sources that the feature cited. In other words, the AI produced statements that its own references did not support.

How did the court reason? According to the report, the court treated AI Overviews as Google's own content and not a neutral pointer to third-party material, so Google bears direct liability. It drew a line between relaying content and authoring it. The AI rewrote and evaluated sources, in the words quoted in the report, in its own words and according to its own structure. The court also rejected the argument that links to sources allow users to verify the summaries, reasoning that citations do not transfer responsibility for the statement. It also noted that Google could have compared its output against the sources and did not.

What is the legal status? The report stresses that the ruling is a temporary injunction, not a final judgment. It does not bind US courts. The analysis argues that it previews a question that American courts may face, since Section 230 of US law protects hosting or transmitting others' content, not authoring it. The United States has no AI-specific liability statute, so existing defamation, negligence and consumer protection law would apply. Platforms may also redesign products to reduce exposure, which would reach users everywhere. Those are the analysts' arguments and not legal conclusions, and the report is not a prediction of how any court would rule.

Why does it matter for insurers? Insurers and brokers increasingly use generative tools to answer customer questions, summarise policies, assess eligibility and support claims decisions. The report says the risk is highest when such a tool makes checkable factual claims about named people or businesses and presents the output as a final answer. Examples it gives include policyholder coverage summaries, eligibility decisions, claims settlement recommendations and underwriting risk assessments. A chatbot that tells a customer they are covered when they are not, or that wrongly says a claim was fraudulent, could create legal and regulatory problems. The report quotes Forrester analyst Nikhil Lai as saying that AI outputs must be legally defensible, not just helpful.

What does the report recommend? It suggests three steps. First, make an inventory of every place where a generative system makes factual claims about third parties or customers, separating verbatim retrieval from synthesis. Second, ground customer-facing outputs in source material and keep logs that link sources to statements, so that a company can show where a statement came from. Third, review vendor contracts to clarify who bears liability when a third-party model produces false statements under the insurer's brand. These steps are a practical checklist and would be useful even if the Munich ruling were overturned.

What are the limits of the lesson? The case concerns a search engine and publishers, not insurers, and German law differs from the law elsewhere. The injunction is temporary. Insurers have regulatory duties, including fair treatment of customers, that exist whatever the court rules. Insurance regulators in Europe and the US are also developing guidance on AI use. A related development on this site is Marsh's report on AI exclusions, which shows that insurers are also deciding how to cover the risks that AI creates for their own customers.

What should customers know? People who use an insurer's chatbot should treat it as a guide and check important points, such as coverage and claims decisions, against the policy wording or a human adviser. Keeping a record of what a chatbot said can be helpful if a dispute arises.

The broader point is about responsibility. A company that deploys an AI tool under its own name is likely to be seen as speaking through that tool, whatever the technology behind it. Insurers already face rules on what they may say to customers, and generative AI makes it easier to say something wrong at scale, which is why controls, logging and human review matter.

This article summarises a published analysis and is not legal advice.