AI Search Liability
An AI Search Answer Isn't the Same as a Link — a Munich Court Reportedly Said So Too.
A German court reportedly held Google directly liable for false, defamatory claims its AI Overviews generated — treating the AI-written summary as Google's own content, not a neutral display of search links. Here's the reported case, the general liability framing behind it, and a practical way to verify an AI claim before you repeat it.
Read this first
Reported facts, general framing, not legal advice
Everything about the Munich case on this page is drawn from public reporting, not the court file — and it's a preliminary injunction under appeal, not a final ruling. The general hosting-vs-authored-content distinction is a framing discussed by commentators, not a settled rule for every AI product or jurisdiction. If an AI answer has actually harmed you or your business, talk to a lawyer — this page is not that.
See the difference
What was reportedly ruled, the general legal distinction behind it, and how to verify a claim before you repeat it.
Why this matters if you rely on AI search answers
- Business owners tracking their name online
- An AI-generated answer engine could summarize your company inaccurately, and you'd only find out from a customer.
- Cross-check a claim about your business across independent models before treating any single AI answer as fact.
- Researchers and journalists
- A synthesized AI answer reads as authoritative but may misattribute a claim to a source that never made it.
- Sourced answers with openable citations let you verify the claim against the actual page before repeating it.
- Anyone quoting an AI search result
- You want to know whether an AI-generated summary and a plain list of search links carry the same legal weight.
- This page explains the general framing courts have started to draw, in plain terms — not as legal advice.
A practical response workflow
- 01
Don't repeat an unverified AI claim as fact
A synthesized answer is not the same as a linked search result — treat a surprising or damaging claim as unverified until you check the source.
- 02
Get a sourced answer with openable citations
@vustSearchBot returns numbered links you can open, so you can check whether a claim actually appears on the cited page.
- 03
Cross-check a contested claim across models
The Council of Sages in @vustbot asks Claude, Grok and Gemini the same question independently and shows where they disagree — useful when one AI answer looks off.
Verify before you repeat it
Open @vustSearchBot for a sourced answer with openable citations before treating any single AI summary as settled fact.
What this page is — and isn't
Informational, not legal advice
This page summarizes publicly reported facts about a court decision and a general legal framing. It is not legal advice for any specific situation — talk to a lawyer if an AI-generated answer has actually harmed you or your business.
The ruling is a preliminary injunction, not a final judgment
As reported, Google has said it will appeal, and the case addresses specific false statements about two named plaintiffs — not a blanket rule for every AI Overview or every AI search product.
We report what's public, not court filings we haven't seen
Case facts below are drawn from public reporting, not the underlying court file. Details of the injunction's exact wording and its international enforcement reach are described by commentators, not independently confirmed by us.
Frequently asked questions
What did the Munich court actually rule?
According to public reporting, a preliminary injunction issued 28 May 2026 (reported as Landgericht München I, case 26 O 869/26) found Google's AI Overviews feature had generated false claims linking two Munich publishers to scams and subscription traps, and treated those AI-generated statements as Google's own authored content rather than a neutral display of search-result links. It ordered Google to stop repeating the specific false statements and reportedly assigned Google roughly 80% of court costs. We're relying on public reporting, not the court file itself.
Is this a final ruling?
No — as reported, it's a preliminary injunction, and Google has publicly said it disagrees and intends to appeal. It also addresses specific false statements about two named plaintiffs, not a blanket rule for every AI Overview or every AI search product.
Does this mean every AI search engine is now legally liable for wrong answers?
No — this page describes one reported case and a general legal distinction commentators have drawn from it (linking to content vs. authoring new content), not a settled international rule that automatically applies to every product, jurisdiction, or kind of AI-generated answer. Treat the general framing as background, not as a verdict on any other product.
What should I do if an AI search result says something false about my business?
This isn't legal advice, but a reasonable first step is documentation and verification: capture the claim, check whether the cited source actually supports it, and get a second opinion by cross-checking the claim in another tool. If the claim is damaging, talk to a lawyer familiar with defamation law in your jurisdiction — a search or AI tool can't tell you your legal options.
How does @vustSearchBot help with this?
Every @vustSearchBot answer carries numbered, openable citations, so you can check whether a claim actually appears on the page it's attributed to before repeating it. For a claim worth stress-testing further, the Council of Sages in @vustbot asks Claude, Grok and Gemini the same question independently and shows where they disagree.
Does a disclaimer like "AI may make mistakes" protect a company from this kind of ruling?
According to reports on this case, a generic disclaimer was reportedly found insufficient on its own to excuse a specific, confidently phrased, false factual claim about a named party — but whether that reasoning holds on appeal or elsewhere isn't settled. It's a reason to treat disclaimers as a starting point, not a complete shield, when you're deciding whether to trust or repeat an AI-generated claim.
Ready when you are
An AI summary isn't the same as a link you can check.
Get sourced, citable answers from @vustSearchBot, and cross-check contested claims with the Council of Sages in @vustbot.