German court says Google can be liable for false AI Overviews
A preliminary injunction found Google responsible for false AI summary claims about two publishers, limiting a key defense for AI search tools.
By Maya Lindqvist · Senior Technology Correspondent
3 min read
A German court has held Google responsible for false statements produced by its AI Overviews feature, issuing a preliminary injunction that bars the company from repeating disputed claims about two publishers. The ruling matters because the court treated the AI summary as Google’s own output rather than a neutral display of third-party search results.
According to Ars Technica, the case involved two publishers who said Google’s AI-generated summaries wrongly associated them with scams and questionable business practices. The Decoder first flagged the ruling, which said Google did not fix the misleading output after the publishers sent a cease-and-desist letter earlier this year.
The court rejected Google’s argument that users know AI answers can be wrong and should check them against other material, according to a Google-translated version of the ruling cited by Ars Technica. The court said the AI Overview made “independent, new, and substantive statements” based on Google’s reading of web links, rather than merely listing external pages.
That distinction was central to the order. Traditional search engines have often avoided direct liability when they point users to potentially harmful material because they are organizing third-party information at scale. The German court said AI summaries are an extra feature that users do not need in order to search the web, according to Ars Technica’s account of the ruling.
Why the court treated AI summaries differently
The court found that the disputed AI Overview included statements that did not appear in the underlying search results, Ars Technica reported. It also said only Google could correct the algorithm and the generated answers shown in AI Overviews, while the affected publishers could not resolve the problem by suing a third-party website.
In weighing the dispute, the court found the publishers’ interest in removing false information stronger than Google’s commercial speech interests, according to the translated ruling cited by Ars Technica. The court described the false AI outputs as tied mainly to Google’s commercial activity and said such statements could influence public opinion.
The decision could affect how courts assess liability for AI-generated answers, especially when chatbots or search summaries paraphrase, misstate or misattribute material from online sources. Ars Technica reported that AI companies have previously leaned on warnings about possible inaccuracies as a defense against claims over false outputs.
Google says the order is not final
Google has not said whether it will appeal. A company spokesperson told Ars Technica that Google works on AI Overview quality so most responses provide accurate information and said the feature is designed to reflect information available on the web. The spokesperson said Google is reviewing the decision and noted that it is not final.
The ruling also comes as researchers and publishers scrutinize how users interact with AI search results. The Decoder cited a Pew Research Center survey from July 2025 finding that many Google users do not click source links when an AI summary appears. It also pointed to a New York Times analysis from May that found Google AI Overviews using Gemini 3 were inaccurate about 9 percent of the time and included inaccurate source links about 56 percent of the time.
The preliminary injunction applies to the false claims at issue in the case. Its broader impact will depend on whether the ruling survives further review and whether other courts adopt similar reasoning for AI-generated search answers.
This story draws on original reporting from Ars Technica.