Artificial Intelligence

Munich Court Says Google Liable for 'AI Overviews'

Is this a defective product deserving liability or a platform that shouldn't be held responsible for its output?

JUN 11·Updated JUN 15

Munich Court Says Google Liable for 'AI Overviews'

Story

Above: The Google Gemini logo displayed on a smartphone and computer screen in Tunis, Tunisia, on May 24. Image credit: Imen Ben Youssef/Hans Lucas/AFP/Getty Images

The Facts

The Regional Court of Munich issued an injunction on May 28 against Google, barring it from repeating false claims about two Munich-based publishers after its AI Overviews linked them to scams and questionable business practices.
India Today
THE DECODER
It classified Google as a direct infringer, ruling that AI Overviews generate "independent, new, and substantive statements" by combining third-party content — unlike traditional search results, which only direct users to websites.
THE DECODER
India Today
Verge
The court found the AI had mixed up information about unrelated companies, creating connections that did not appear in any linked sources. Google was ordered to cover 80% of legal costs, with each publisher paying 10%.
Times Now
Engadget
The court rejected Google's argument that users could verify AI summaries by checking linked sources, noting that studies show users are far less likely to click on source links when an AI Overview appears in search results.
THE DECODER
Mezha
A Google spokesperson said the company was "carefully reviewing this decision, which is not yet final," noting the ruling is part of a preliminary proceeding and that it invests in AI Overviews quality to ensure accurate responses.
Crypto Briefing
Android Authority
An analysis by AI startup Oumi found Google's AI Overviews answered accurately 91% of the time, but 56% of correct answers could not be verified through the linked sources, an issue the Munich court addressed directly in its ruling.
Telecompaper
THE DECODER

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Narrative A

Google's AI Overviews are original documents the company creates, and a Munich court got that right. The AI invents claims out of thin air, confidently presents false information as fact, and reaches 2 billion users monthly, meaning billions of incorrect queries every year. Letting Google dodge liability by calling it a search engine would be a legal fiction that lets a defective product keep harming real people.

Owen Gregorian on X

Whatever Blog

Narrative B

The Munich ruling is a troubling legal overreach that could spread through international courts and distort how AI-powered search is regulated globally. Holding Google liable for AI-generated content sets a dangerous precedent that misclassifies algorithmic output as editorial speech — a framework that would be unworkable for generative AI at scale. This signals that activist courts are willing to dismantle the platform liability protections underpinning an open web.

Gary Marcus on X

PC Gamer