A regional court in Munich ruled in May 2026 that Google can be held directly liable when its AI Overviews state false, damaging claims about a company, because the court found that an AI Overview is Google's own written statement, not a list of search results. A Berlin court reached the opposite conclusion a few weeks later in a different case, and Google is appealing the Munich ruling. Nothing here is settled law yet, in Germany or anywhere else, so the practical answer for a brand worried about AI misinformation has not changed: don't wait for a court to fix it for you. Here is what the rulings actually say and why they matter anyway.
What did the Munich court actually rule?
That Google is directly responsible for the content of its own AI Overviews, not merely a neutral platform displaying links.
The Regional Court of Munich I, in case 26 O 869/26, issued a preliminary injunction on May 28, 2026, after an oral hearing on April 23, 2026, according to The Decoder's reporting. The case was brought by two Munich-based publishing companies after Google's AI Overviews, in response to searches for their company names, generated text connecting them to a "Betrugsmasche," a fraud scheme. The banned claims included that the companies lured customers into subscription traps, invoiced them for "company directory listings" or "Premium Gold Packages" after calls that never took place, kept demanding payment after customers had already paid, and changed names and URLs to dodge identification. The court also found that some of the AI-generated connections between the plaintiffs and other named companies did not appear in any of the sources the Overview cited.
Why did the court say Google is liable, not just displaying results?
Because it found that an AI Overview writes its own sentences, in its own structure, rather than reproducing what a source actually said.
The court's reasoning, as reported by The Decoder, rests on a distinction between showing links and generating prose. A traditional search result points to someone else's words. An AI Overview, the court held, produces "independent, new, and substantive statements" that Google itself authored, even when it draws on other pages for raw material. That matters because German case law from the Federal Court of Justice (BGH) has long given search engines limited liability for the third-party content they merely index and display. The Munich court ruled that protection does not extend to a feature that writes its own answer. The court also rejected Google's argument that users could check the underlying sources themselves, and it rejected the idea that Google could claim host-provider protections under the EU's Digital Services Act, reasoning that the companies would otherwise have no effective way to get a false, damaging statement corrected.
Is this now settled across Germany, or even inside this one case?
No, on both counts.
The Munich ruling is a preliminary injunction, not a final judgment, and Google has said it will appeal. A Google spokesperson described the case as focused on "specific and narrow errors, not the foundational way AI Overviews displays web content," according to The Decoder's follow-up report. The same report notes that a Berlin court reached the opposite conclusion in a separate case a few weeks later, treating AI Overviews as another form of search result and holding that Google carries only the limited, indirect liability a search engine normally has. Google is expected to cite that Berlin ruling in its appeal. Two courts in the same country, applying the same underlying law to the same product, landed in different places within a month of each other. That is not a sign of a settled legal standard. It is a sign that the legal system is still working out how to categorize a feature that didn't exist in its current form a few years ago.
Does a ruling like this actually help a company that's been misrepresented?
Only the two plaintiffs in this specific case, and only for now, since the injunction can still be overturned on appeal.
It is worth being precise about what the Munich ruling does and does not establish. It does not create a general right for any company anywhere to sue Google over an AI Overview. It is a German civil court applying German personality-rights law to a specific set of facts, and it conflicts with another German ruling from the same period. For a company outside Germany, or dealing with a different AI platform entirely, this case is not a template, it's a data point about how one legal system is wrestling with a genuinely new question: who is the author when software writes a sentence by combining and paraphrasing other people's words.
What should a brand actually take from this, short of legal advice?
That AI-generated claims about your company are not a legal gray area you can rely on; they're an operational problem you still have to manage directly.
Three things follow from the facts of the case, independent of how the appeal goes. First, Google's own feedback mechanisms inside Search and its AI products remain the fastest way to flag a specific wrong answer, the same point covered in what to do when an AI assistant gets a fact about your company wrong. Second, the Munich plaintiffs won in part because they could show the AI's claims didn't match any of the sources it cited, which means keeping a dated record of what an AI tool says about you, with screenshots, is the kind of evidence that actually holds up if you ever do need to escalate a complaint. Third, the underlying fix is almost always upstream: an AI system generates a false claim because something in its training or retrieval data supports it, however loosely, which is why the sites an AI platform actually trusts and cites matter more to what gets said about you than any single correction ever will.
Why does this matter even though nothing is legally settled?
Because it's the clearest public evidence yet that courts are treating AI-generated answers as the platform's own speech, not a neutral reflection of the web.
Whatever happens on appeal, the Munich court's framing, that an AI Overview is "Google's own words" rather than a display of someone else's, is a real argument now on the record, made by a court, in a case Google is actively contesting rather than ignoring. That is a meaningfully different posture than a feedback button buried in a settings menu. It also lines up with a broader pattern: AI platforms are facing more scrutiny over what their answers say and where that content comes from, at the same time that the mix of sources those platforms actually cite keeps shifting underneath everyone trying to track it. None of that gives a brand a legal shortcut. It does mean that getting your own facts straight, in places an AI system actually reads, is doing more work than it used to.
The direct answer, for anyone who read this far
A Munich court ruled Google can be directly liable for false, damaging claims its AI Overviews generate, reasoning that the Overview is Google's own authored statement rather than a passive list of links. A Berlin court reached the opposite conclusion weeks later, and Google is appealing the Munich decision, so the law here is unresolved, not settled, and not something to plan around. What's actually useful from the case is the evidence standard it reveals: the plaintiffs won in part because they could show the claims didn't trace back to any real source. Keep that kind of record for your own company, report specific wrong answers through the platform's own tools, and put effort into the source material AI systems actually read, because that is still the only lever you control regardless of how any appeal turns out.
If an AI assistant has already published something false or damaging about your company, talk to us about fixing it at the source.
