A German court has ruled that Google is directly responsible for incorrect information presented by its AI Overviews platform, according to a report from The decoder. The country has laws in place that protect search engine operators from liability, but the court ruled that this did not apply to AI previews. It classified Google as a direct infringer because the AI Overview is its own content and not just a list of search results.
It all started when the company’s AI summary algorithm spread false information about two Munich-based publishers. Publishers were linked to scams, subscription traps, and questionable business practices through certain search queries. The court says the AI mixed information about completely separate companies, making connections that did not appear in any linked sources and did not actually exist.
The publishers sent Google a cease-and-desist letter, but they say the company has not responded appropriately. The Munich Regional Court imposed a temporary injunction on Google prohibiting it from spreading false information about the two companies involved in the case.
The decision places the blame for any factual errors on Google, as the AI Overview rewrites the information “in its own words and according to its own structure.” In this case, the overview confidently suggested that one of the publishers was “known for questionable business practices” and had built its own structure with a summary, red flags for these questionable practices, and tips for users. The problem, again, is that AI Overview was actually extracting information about another entity. He even made up claims out of thin air that weren’t noted in search results.
Simply put, the Google search engine would not be directly responsible because it only makes third-party content findable. Google, the AI Overview operator, is responsible because these overviews create “independent, novel and substantive statements” by evaluating and combining content from third-party websites. That’s pretty much what I do as part of my job, and I would certainly be responsible if I made up a bunch of defamatory things about a few publishing houses.
Google argued during the hearing that users could check the linked sources to verify whether the AI summary was correct. He also said these users know “not to blindly trust AI-generated information.” That’s a pretty remarkable statement considering how quickly these glimpses of Gemini were thrust upon us. Studies suggest that only 1% of users click on source links after reading one of these previews.
“Google users are less likely to click on links when an AI summary appears in results” according to a new study from @pewresearch: pic.twitter.com/812XtjqLN5
— Bob Pickard (@BobPickard) July 22, 2025
How serious is the AI Overview’s factual accuracy problem? It depends on your point of view. A recent study reported by The New York Times suggested that the preview is wrong about nine percent of the time. This seems like a pretty decent metric until you realize how busy the platform is. Google recently touted that 2 billion people interact with AI insights every month. This represents 24 billion people each year. Some simple math suggests that this translates to over 2 billion incorrect queries each year. This is an extremely conservative estimate, given the ubiquity of the platform and the number of Google searches (16.5 billion) performed each day.
However, factual accuracy is not the only issue here. One study noted that these previews also pose a supply problem. This analysis found that 56% of correct answers could not be supported by the linked source. In other words, there is no real way for users to actually verify the work of the AI.
By the way, there is a way for users to “miss” all this magical AI search technology. Simply insert “-ai” at the start of any Google query. You are welcome.
