German Courts Now Treat Google's AI Answers as Google's Own Speech

A German court has ordered Google to stop publishing three versions of an AI Mode answer about a local doctor — backed by fines of up to 250,000 euros per violation — even though every fact inside the answer was true. The facts simply belonged to a different man.

In a decision dated Sept. 24, 2026, the Regional Court of Aachen ruled on an emergency basis, without first hearing from Google, after the platform declined to act. The doctor, a former chief physician and professor of neuroradiology who left his post in late June, had sent Google a detailed explanation in late August. In mid-September Google said it would take no action; a formal legal demand sent the same day also went unanswered. Google can still challenge the order.

What did the AI answer actually say?

When users searched for why the doctor left, AI Mode supplied a serious criminal allegation. Follow-up queries using terms like "suspension," "termination," and "court proceedings" produced multi-paragraph accounts of a legal aftermath. The proceedings were real — but they involved another former university professor, and the court confirmed the mix-up from its own records. The takeaway: a platform can be liable for an answer assembled entirely from accurate facts if the answer fuses two entities into one.

Where does the liability logic come from?

Aachen relied on a May ruling by the Regional Court of Munich, which held Google directly liable for AI Overviews that tied two publishers to fraud and subscription traps using claims absent from every linked source. Google argued users could verify the sources themselves. Munich rejected that, finding the overview read as a self-contained statement, and described Google's AI output as "independent, new, and substantive statements" built by combining third-party content. Aachen adopted that language directly. Google appealed the Munich ruling, then — according to The Decoder's reporting — settled before any higher court weighed in.

Why does a Berlin court see it differently?

In early June, a Berlin court reached the opposite conclusion in a trademark dispute. A perfume company objected to AI Overviews that named its brands next to links selling cheaper imitations. Berlin treated AI Overviews as a new format of search result bundling other sites' information, not as Google speaking for itself, and found no infringement. The distinction is real: Berlin's summary accurately repeated what sources said, while Munich and Aachen involved invented connections. Even so, no higher court has ruled, because the one case headed there settled.

How common is this failure mode?

The pattern repeats. In Aachen, two men shared the attributes of professor and departure from a post. In Munich, the AI blended publishers with companies that, per sworn statements, had no connection to them. In 2024, Microsoft's Copilot described a German court reporter as the perpetrator of crimes he had covered, because his name sat beside those crimes in his own bylines. Duane Forrester, writing on Duane Forrester Decodes, calls this "entity conflation" — the system merges two distinct people or companies into one and speaks about the result with confidence. Any business sharing a name or attribute cluster with a neighbor is a candidate, and most have never checked which neighbors those are.

What changes for marketers?

The risk is not confined to Germany. Platforms rarely maintain separate answer pipelines per jurisdiction, and the Munich reasoning rests on how the product works, not on who built it. Pew Research found users clicked a link inside an AI summary on just 1% of searches where one appeared, versus 15% clicking traditional results when no summary appeared — which is precisely why courts treat the answer as the platform's speech. Forrester's practical guidance: query your brand with role, location and event terms; publish disambiguating facts in plain sentences; deploy Schema.org Organization and Person markup with sameAs links; publish your own account of material changes; and document every erroneous answer, query and correction request — both German cases turned on documented requests that went unanswered.

The open question is whether precedent ever forms: if platforms keep settling quietly, liability stays a manageable cost rather than a design constraint, and the pressure to fix conflation weakens.