German Court: Google Liable for AI Summaries

Geo-Specific , Litigation , Standards, Regulations & Compliance German Court: Google Liable for AI Summaries Court Says Google Is Responsible for Content Generated by Its Machines David Meyer • June 12, 2026     Credit Eligible Get Permission Image: Marco Lazzarini/Shutterstock A German court dealt a blow to Google artificial intelligence-infused search by deciding that the Big Tech firm is liable for defamatory statements in the AI summaries that increasingly dominate its results. See Also: Cybercrime - Prepare | Protect | Respond A regional court in Munich granted a preliminary injunction in late May to local publishing house Verlagshaus24 and one of its subsidiaries, demanding that Google immediately stop its AI Overviews summaries from generating false claims about their business practices. This feature purports to only summarize publicly-available content, but it often hallucinates facts - and in this case, it misinterpreted third-party content and said the plaintiffs were the targets of accusations that had actually been made against another publisher. The injunction, which is effective across the European Union, could have far-reaching consequences because it treats the content of AI summaries differently than traditional search results. With standard search, providers such as Google are generally immune from liability in the EU over assertions in linked content. In this case, the court said the AI summary was "by no means based solely on third-party content." The AI drew its own conclusions, meaning Google was responsible for it. "The Munich court has sent a clear and important signal that providers cannot hide themselves behind AI in the event of misrepresentations," Bernhard Buchner, a lawyer for the plaintiffs, told ISMG on Thursday. "New and innovative developments must always comply with the existing legal framework, and individual rights cannot be sacrificed for the sake of a trend." A Google spokesperson insisted that the company invests "deeply in the quality of AI Overviews to ensure that the overwhelming majority of responses provide accurate information, and they are designed to reflect the information that exists on the web," and said it was "carefully reviewing this decision, which is not yet final." The dispute began in early February, when the publishers sent Google a cease-and-desist letter because AI Overviews was telling people they have a reputation for engaging in scams and dubious business practices. These included "subscription traps" where customers are lured into unknowingly buying subscriptions, invoices for services on the basis of phone calls that never took place, demands for payments that had already been made and a failure to unlock digital content that had been paid for. Google's AI feature continued to make some of these allegations after the first warning letter was sent - currently, it explicitly states that Verlagshaus24 is not operating fraudulent schemes. The Munich court sided with the plaintiffs on most of their claims, saying their corporate personality rights had been damaged. It rejected a relatively minor claim about AI Overviews associating the companies with a debt collection agency - judges decided this was not a particularly damaging assertion - and another about selling subscriptions "in a dubious manner," since the evidence didn't show the AI summaries clearly making this assertion. The court said precedent set by the German Federal Court of Justice in earlier rulings limited liability for search engine providers because there is no way search engines could proactively verify the legality of every piece of content they index. When it comes to features like AI Overviews, it argued, that logic does not hold. "What is displayed is precisely not merely search results from websites with which [Google] has no contact and whose content it cannot readily verify itself. Rather, independent, new and substantive statements are made, based on an evaluation and combination of the content of various third-party websites," the court wrote. "The content of these new statements, made by the AI offered by [Google] itself, is entirely capable of being reviewed by [Google] - if need be through appropriate control mechanisms - at least by comparing the underlying third-party websites with [Google's] own statements based on them." AI Overviews are "by no means indispensable" for using the web, the court said. And, in a rebuke to the use of generative AI in search, it also expressed little patience for Google's argument that users can use the links in the summaries to verify that the underlying content supports the claims being made. "The possibility that a statement can be falsified through further research does not, as a rule, exempt one from liability for that statement," the ruling reads. "The very utility of the function as asserted by [Google] would be significantly diminished if the 'AI Overview' were generally to be treated as unreliable and all of the displayed links had to be checked independently in each case." Pew Research found that barely 1% of Google Search users click through to cited links within AI Overviews. The Munich court's ruling follows another, made by a Frankfurt court in September 2025, which established that a search engine provider could theoretically be held liable for false information in its AI summaries - although in that case, which involved competition law, the plaintiff failed to win the injunction it was after. "The ruling strengthens the position of companies whose reputations are harmed by AI-generated search results," IT lawyer Stefan Lutz wrote in a Thursday analysis of the Munich decision. "Previously, it was often unclear whether affected parties could take action against the original sources, the platform operators, or anyone at all when the actual misinformation arose only through the AI summary. The Munich I Regional Court has now expressly closed this gap in legal protection. The judges pointed out that, otherwise, affected parties would often lack effective legal recourse, as the contested statement does not originate from the linked third parties but is generated by the AI itself." Lutz added that the ruling "increases the pressure on providers of generative AI systems to design their models and control mechanisms in a way that minimized the risk of serious misattributions." The European search-engine liability protections that did not apply in this case are comparable to those granted in the United States through Section 230 of the Communications Decency Act, which is broadly seen as having allowed online platforms to flourish without fear of being held responsible for the third-party content. It remains unclear whether AI companies could rely on Section 230 protections for the content generated by their models, but legal experts say recent rulings suggest this is unlikely in cases where the AI itself materially contributes to the content.