The moment Google can't hide behind "it's just search" anymore, the liability landscape shifts. A Munich regional court just proved it.
On June 9, 2026, the Regional Court of Munich ruled that Google is directly liable for false claims made by its AI Overviews. Not indirectly. Not "third-party content." Directly. The AI Overview isn't a search result. It's Google's own statement.
This changes everything. Here's why.
When AI Writes, Someone Owns It
Google's traditional argument (that search engines just point users to third-party content) doesn't work for AI Overviews. The court made this crystal clear. An AI Overview doesn't link to statements that already exist on the web. It synthesizes, rewrites, and interprets information "in its own words and according to its own structure." That makes it original content. And original content has an author. That author is Google.
In the Munich case, Google's AI had confidently declared that two publishing companies were "known for dubious business practices" and linked them to subscription scams. None of those claims appeared in any of the sources the AI cited. The AI made them up by connecting dots that didn't exist. When the publishers sent a cease-and-desist, Google didn't treat it seriously. The court did.
The ruling was brutal: Google had to stop making those false claims, pay 80% of the plaintiffs' legal costs, and confront a simple fact. If your AI can generate false statements, you're liable for them.
Why "Users Can Check" Doesn't Protect You
Google's defense at trial was basically: people should fact-check AI Overviews themselves. Users know not to blindly trust AI, right?
The court rejected this entirely. And it did so with a specific argument that matters for every brand: an AI Overview is "understandable on its own" and contains "a self-contained statement with independently understandable content." Users don't need to click sources to understand what the Overview is saying. They read it, they believe it, and they move on.
Research backs this up. Studies show that users almost never click on the sources linked in AI Overviews. So Google's argument that "people can dig deeper" isn't just legally weak. It's factually absurd. Most users won't dig deeper. They'll just absorb the false claim.
The court also drew a parallel to press law: if a newspaper publishes a false teaser that's understandable on its own, the publisher is liable even if readers never click to the full story. Same logic. Same liability.
The Scale Problem
Google's Gemini 3 model answers correctly about 91% of the time. That sounds good until you do the math. At Google's scale, a 9% error rate means millions of false answers every hour. Some of those false answers will defame companies. Some will make false claims about individuals. Some will falsely connect legitimate businesses to scams or illegal activity.
The Munich court's ruling didn't say Google needs 100% accuracy. It said: when false claims harm someone, you're liable. And with millions of errors daily, the liability exposure compounds.
What makes this worse for brands: the false claim might not even be about your company directly. Google's AI might have mixed your company's name with another company's sketchy history. You didn't create the false connection. Google's algorithm did. But you're the one whose reputation gets damaged.
What This Means for Brands
If you're a CMO or a brand safety leader, this ruling is a wake-up call on three fronts.
First, your brand name is now more vulnerable to false association through AI Overviews. A competitor could seed sketchy content on the web that, when interpreted by an AI, creates a false link to your company. You can't prevent this. You can only sue Google after the damage is done.
Second, Google now has liability incentives it didn't have before. That means Google will be more cautious about what AI Overviews say. But caution takes time. In the interim, false claims will still get published to millions of users before Google takes action.
Third, this ruling will likely spread beyond Germany. The EU's Digital Services Act gives other countries similar grounding. If Germany's regional court can hold Google liable, higher courts elsewhere will have to answer the same question: is an AI Overview a search result, or is it Google's own published statement?
The answer is increasingly obvious. It's Google's statement. Which means Google owns what it publishes.
The Broader Pattern
This isn't just about Google. Any company publishing AI-generated content at scale is now facing the same question. ChatGPT. Claude. Perplexity. If you generate new statements that synthesize third-party information, and those statements are false, are you liable?
For search engines, the answer used to be no. Search engines just indexed and ranked other people's content. But AI Overviews aren't indexing. They're authoring. And authors are responsible for what they publish.
The Munich court also rejected the idea that being a "host provider" shields you from liability. That protection doesn't apply when you're the one creating and publishing the content. Google can't be a neutral platform and an author at the same time. The court decided: you're an author. Act like one.
The Bottom Line
Google's AI Overviews are no longer legally invisible. They're now Google's own published statements, subject to the same defamation and false advertising laws as any other publication. That ruling will ripple through the industry. It will reshape how AI companies approach publishing at scale, and it will force brands to rethink how they monitor and respond to false claims made about them in AI systems they don't control.
The question isn't whether other courts will follow Munich's logic. It's how fast they'll do it.
For more on AI liability and brand risk, see our recent coverage on vendor liability traps in AI adoption and what brands need to know about AI disclosure compliance.