Breaking

Munich Court Finds Google Directly Liable for False AI Overviews Statements

June 13, 2026 at 22:38 EDT

Policy & Regulation
Foundation Models

The Regional Court of Munich (Landgericht München I) issued a preliminary injunction around May 28, 2026, holding Google legally responsible for false statements displayed by its AI-generated summary feature, AI Overviews. The ruling holds that a company that designs, trains, operates, and manages an AI system must assume legal liability for any damages caused by the responses it generates.

May 28, 2026 · Regional Court of Munich I · Case 26 O 869/26

Google Is Liable for What Its AI Says

A German court ruled that false statements produced by Google's AI Overviews are Google's own words — not third-party content — stripping away the intermediary liability shield in a first-of-its-kind precedent.

€250K

Maximum fine per violation if false statements recur

80%

Share of court costs borne by Google

Munich publishers who brought the case

The legal pivot

Where the immunity stops

Search results

Traditional links

SHIELDED — third-party content

Synthesized claim

New, substantial

AI Overview

Generative AI output

NOT SHIELDED — Google's own speech

Because AI Overviews generate "independent, new and substantial statements," the court held that hosting immunity under the EU e-Commerce Directive and Germany's Telemediengesetz does not apply.

The statement at issue

"Yes, [it] is known for dubious business practices and is often perceived as a scam."

— text not found in any linked source, but newly synthesized by the AI from multiple inputs.

The court's reasoning

"Nobody needs AI to search the internet." Google argued users could verify claims themselves; the court rejected this, saying AI Overviews are not essential to using the web — so Google remains the responsible party.

Google's position

Google will appeal, arguing the case concerns specific, narrow errors — not the foundational workings of AI Overviews. The injunction is provisional, ahead of a final ruling.

Ripple effects

Even at a reported ~90% accuracy, Google's enormous search volume could mean tens of millions of erroneous answers. The precedent may extend to any service that synthesizes sources into answers:

Perplexity Microsoft Bing AI OpenAI search

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.