On May 28, 2026, the Munich Regional Court I (Landgericht München I) issued a landmark ruling under case number 26 O 869/26 on the liability of AI search systems. The court classified Google LLC as a direct tortfeasor for false claims generated by its AI-powered feature known as AI Overviews within its search results. Two Munich-based publishing companies had sued Google after the feature falsely linked their names to fraud schemes and dubious business practices. The court prohibited Google from continuing to disseminate the contested content and imposed a potential fine of up to 250,000 euros for each violation. While the judgment is not yet final, it has already attracted significant attention in German legal practice.
The decision is particularly significant because it makes a clear statement about how AI-generated search content must be treated under liability law. Google had attempted to rely on the liability privileges of traditional search engine law – specifically the principle that a search engine merely links to and displays third-party content and therefore bears no responsibility for that content. The Munich court convincingly rejected this argument. The AI feature does not produce a mere index of third-party content; rather, it generates independent, coherent texts that evaluate multiple sources and condense them into a new statement. This independent text generation makes Google the party responsible for the content produced.
The plaintiff publishers found themselves confronted with serious false claims through Google's AI Overviews feature. The AI generated texts that falsely connected the publishers' names with fraud schemes, subscription traps, and dealings with dubious business partners. For the affected companies, this meant significant reputational harm and the concrete risk of revenue losses, as potential customers and business partners encountered the misleading AI summaries as the first search result for the publishers' names. The plaintiffs asserted violations of the corporate right of personality as well as claims under German unfair competition law (UWG). The court granted the claims in full.
Particularly instructive is the nature of the errors produced by the AI. These were not mere typographical mistakes but qualitative misstatements that do not withstand scrutiny of the actual facts. This phenomenon is known in AI research as hallucination – AI language models tend to produce statements with great confidence that are factually incorrect yet grammatically and stylistically plausible. In this case, the AI combined and interpreted content from various sources in such a way as to produce new, false assertions that did not appear in any of the sources used. This source of error is structurally inherent and cannot be fully eliminated through better prompting or improved data quality alone.
Liability law in the digital sphere has traditionally distinguished between different types of service providers. Host providers benefit from liability privileges under the German TDDDG and Article 6 of the Digital Services Act (DSA): they are generally only liable for unlawful third-party content if they are aware of it and fail to remove it promptly. Similar principles apply to search engine operators. The Munich court made clear that the AI Overviews feature does not fall into this category. The AI does not deliver third-party content; it creates its own – and therefore Google must be treated not as a host provider but as a content provider who bears direct responsibility for its own statements.
The court also grounded its reasoning in the personality rights of the affected publishers. The corporate right of personality protects companies against false statements of fact that impair their business reputation. Under German civil law doctrine, anyone who participates in an unlawful impairment of a protected legal interest – even without fault – can be held liable as a tortfeasor. Google, as the operator of the system that generates and feeds the false content into search results, clearly fulfills this requirement. The fact that Google profits economically from the feature and possesses the technical means to control or prevent the system's outputs further reinforces the attribution of responsibility.
It is noteworthy that the Munich court's reasoning diverges from that of the Berlin Regional Court in a parallel case concerning Google AI Overviews. The Berlin court had classified AI Overviews as non-independent content and applied the liability privileges for search engine operators. This diverging case law illustrates that the question of liability for AI-generated search content has not yet been definitively resolved in Germany. It can be expected that the fundamental question will be carried through the appellate courts, potentially all the way to the Federal Court of Justice (BGH) or the European Court of Justice (ECJ). For companies in a similar position to the plaintiff publishers, however, the Munich ruling already provides a solid foundation for legal action today.
The Munich ruling has immediate practical relevance for two groups: companies and individuals harmed by false AI search answers, and businesses that themselves operate AI systems generating comparable texts. For the first group, the message is clear: those confronted with false claims in an AI system do not have to accept this passively. Concrete claims for injunctive relief and damages exist under personality rights and unfair competition law, and they can often be enforced rapidly by way of a preliminary injunction. The Munich ruling significantly strengthens this position by directly and unequivocally affirming the liability of the system operator.
For operators of AI systems offering generative search answers, AI chatbots, or automated summaries, the ruling represents a substantial expansion of their liability exposure. Anyone using an AI system to generate independent textual statements about third parties must expect to be held directly responsible for the content of those statements. Technical quality assurance measures, content moderation, disclaimers, and feedback mechanisms are therefore indispensable not only from an ethical but also from a liability law perspective. Careful assessment is required to determine whether one's own AI services fall within the liability privileges for search engine operators or whether they qualify as independent content creation – a distinction that can be decisive for legal responsibility in a dispute.
Companies affected by false AI search answers should first document the specific statements thoroughly – using screenshots with date and time stamps, ideally reproduced across different devices and browsers to demonstrate reproducibility. The next step is to seek legal advice to assess whether the statements are indeed false and unlawful and whether claims for injunctive relief or damages can be pursued. In many cases, a preliminary injunction is the most effective instrument for achieving rapid relief. In parallel, a direct request to the AI system operator for out-of-court removal of the false content should be submitted – since providers like Google now have good reason, in the wake of the Munich ruling, to respond to such requests promptly. For AI system operators, an immediate legal review of their liability exposure and the implementation of robust mechanisms to prevent, detect, and correct AI hallucinations are strongly recommended.
The Munich Regional Court I ruling of May 28, 2026 is an important milestone in German jurisprudence on liability for AI-generated content. It makes clear that the traditional liability privileges of internet law cannot be reflexively applied to new AI features – the concrete technical functionality of the system is decisive for determining liability. AI systems that generate independent content are not equivalent to classical search engines, and those who operate such systems must answer for their outputs. At the same time, the ruling opens a clear legal path for companies and individuals harmed by AI hallucinations to pursue their claims. Given the rapidly growing prevalence of generative AI applications across all areas of life, AI liability will remain one of the central topics of German and European IT law in the years to come.
Do you have questions about liability for AI-generated content or the legal enforcement of claims against AI systems? The lawyers at HUFELD PartGmbB are here to provide comprehensive advice. Get in touch now.
Are you looking for a specific topic?
.png)
.png)
.png)