The disclaimer didn’t hold.
According to reporting by The Decoder, the Landgericht München (Munich Regional Court) issued a preliminary injunction finding Google directly liable for defamatory statements generated by its AI Overviews search feature. The court reportedly found that Google failed to act after receiving a cease-and-desist from affected publishers – and then rejected Google’s argument that users understand AI outputs require verification.
That last part is the one compliance teams need to read carefully.
The ruling, as reported, treats AI Overview outputs as Google’s own statements rather than indexed links to third-party content. The court’s reasoning, as reported, holds that because Google controls the model’s logic and training data, the generated output constitutes Google’s own statement, not a referral to what someone else wrote. Case number has not been confirmed in English-language sources; the ruling is a preliminary injunction, not a final judgment, and an appeal remains possible.
Who This Affects
Why it matters
Every major AI platform operating in the EU uses some version of the disclaimer Google used: “AI can make mistakes, check important info.” According to reporting on the ruling, the Munich court found that language insufficient. If that holding survives appeal and influences broader EU jurisprudence, the disclaimer layer that legal teams have been counting as a liability buffer is thinner than assumed.
The ruling sits inside a pattern that’s been building quickly. The registry shows the Penguin Random House v. OpenAI case also landing in Munich, alongside CNN v. Perplexity on content extraction, three AI content liability cases in roughly 30 days, across defamation and copyright theories. German courts are becoming an active venue for AI liability litigation, and the EU’s AI Act’s Article 52 transparency obligations don’t obviously resolve the defamation question the Munich court addressed. Transparency requirements and defamation liability operate on separate legal tracks.
Context
The EU’s Digital Services Act establishes platform liability rules for illegal content, but its primary focus is on hosted third-party content, not AI-generated statements the platform produces itself. The Munich court’s framing of Google as the originator of the output, not a distributor of third-party content, is precisely the distinction that makes this ruling potentially significant beyond Germany. Whether EU-wide jurisprudence follows this logic is an open question, but it’s no longer a purely theoretical one.
Evidence
What to watch
Google’s response to the preliminary injunction matters: does it modify AI Overviews disclaimers, block the outputs at issue, or appeal? The cease-and-desist failure is already in the record, which limits certain defensive arguments. The next milestone is whether the injunction becomes a final judgment and whether other affected parties file in the same jurisdiction.
The real question is whether this is a one-court ruling or the opening of a liability theory that spreads. Three AI content cases in 30 days across two jurisdictions suggests the latter. Compliance teams running AI-generated content products in the EU, chatbots, search summaries, RAG-powered tools, should be asking a specific question: if a court treats our outputs as our statements, what does our disclaimer actually protect against?