When an AI tool says something false about a real company, who pays? For a long time the tech industry's answer was, in effect, nobody: the machine generated it, the sources were linked, the user could check. In June a court in Munich rejected that answer for Google. The ruling is preliminary and Germany is not the US, but the logic is spreading, and it applies to every business that lets an AI speak in its name.
What happened
According to The Decoder, the Regional Court of Munich issued a temporary injunction on May 28, 2026 that barred Google from repeating false claims about two Munich publishers in its AI Overviews, the AI-written answers at the top of search results. For certain searches, the summaries had linked the publishers to scams and subscription traps. The outlet reports that the AI mixed them up with other companies and drew connections none of the cited sources made. The publishers had sent a cease-and-desist letter first.
The court treated Google as the direct author. Older German rulings limit a search engine's liability because it only points to other people's pages. The court said AI Overviews are different because they generate new statements that Google alone controls. It rejected the argument that users can click through and check, and held that Google could not lean on the hosting protections in the EU's Digital Services Act. Google was ordered to cover 80 percent of the legal costs. A Google spokesperson told The Decoder the company is reviewing a decision that is not yet final, and said the vast majority of AI Overviews are accurate.
The Next Web adds context. It notes this is a preliminary injunction, not binding precedent, and that Google can appeal. It compared the court's reasoning to press law, where a misleading teaser can be actionable even if few readers see the full story. And it points out that if the reasoning holds, it could reach other AI answer engines too.
A similar fight is playing out in the US. Minnesota Lawyer reported in 2025 that Wolf River Electric, a Minnesota solar installer, sued Google for defamation after an AI Overview allegedly said the state attorney general had sued the company. The company says no such lawsuit existed, that none of the four sources the AI cited said so, and that it lost customer contracts as a result. Those are allegations, and the case has not been decided on the merits in any reporting we found.
Why it matters for your business
You are not Google. But you may have a chatbot on your website, an AI writing product descriptions, or an assistant drafting customer emails. This is not legal advice, but the Munich reasoning maps neatly onto those uses.
If your AI says it, you said it. The court's core point was that a company that deploys an AI to generate its own statements owns those statements. That is how most customers, and probably most judges, will see your chatbot too.
"The sources were linked" is not a shield. Studies cited by The Next Web suggest very few users click a source from an AI Overview. A false claim sitting in a confident summary does damage before anyone checks.
AI can hurt you from the outside, too. The Wolf River story is a warning for every business owner. An AI answer engine can attach someone else's scandal to your name. You need to know what these tools say about you.
What we'd do
- Treat customer-facing AI like an employee. Give it a narrow job, approved facts to work from and clear limits. No answers about competitors, legal matters or other people.
- Ground it in your own content. Tools that answer only from your verified documents tend to stay much closer to the facts than open-ended chat.
- Log and review. Keep transcripts of what your bot tells customers and review a sample regularly.
- Have a correction path. If the bot gets something wrong, fix it fast and tell the customer. Speed matters in any dispute.
- Search yourself in AI tools. Ask the major AI assistants and search engines about your business every month. If one states something false, document it and use the platform's feedback and legal channels.
- Check your vendor contract for who bears liability for false output.
Our view
Google's position has real weight: AI summaries are built from public sources, they are usually right, and making providers liable for every slip could make useful features too risky to offer. But the Munich court asked the right question. When a company chooses to speak through an AI, and profits from it, the company owns the words. We expect courts in more places to land in the same spot. Businesses should not wait for that. Put your AI on a short leash, check what others' AI says about you, and never rely on "the computer said it" as a defense.
Want a system like this in your business?
A discovery conversation is where we map it: what to automate first, where the human checkpoints go, and what it would take.
Book a discovery conversation →