German Court Holds Google Responsible for False Statements in AI Overviews

German courts have continued to find that Google bears responsibility for false information produced by its search feature “AI Overviews.” The Decoder reported on October 18 (local time) that the Aachen Regional Court applied reasoning holding Google liable. The case involved an AI answer that wrongly attributed a criminal offense committed by someone else to a doctor. Earlier, a Munich court had also barred specific AI Overviews content from being published.

Whose words does an AI-written summary represent?

The central question is whose words an AI-written answer counts as, when that answer is assembled from search information. AI Overviews uses AI to summarize search results and displays them alongside links to relevant web pages. In the Munich case, Google argued that it is merely a search provider surfacing third-party information, and that it had not adopted the content as its own assertion.

Munich Regional Court I found that AI Overviews goes beyond simply displaying search results or links. If the system evaluates what it finds on its own and summarizes it in its own sentences, the court reasoned, the result counts as Google’s own statement. On May 28, 2026, the court ordered that content linking two publishers to fraud and dishonest business practices no longer be published or distributed. The publishers argued that the content violated their corporate personality rights.

Will other courts see it the same way?

In the Aachen case, too, the reasoning that treats AI answers as Google’s own content was applied, according to The Decoder. The target shifted from publishers to a doctor, but the underlying issue – who answers for an AI-generated description – is the same. The outlet also reported that Google appealed the Munich ruling and then settled the case. At the time the Munich court issued its decision, that decision was not yet final.

Other German cases have gone Google’s way. In a 2025 case, the Frankfurt Regional Court noted the principle that Google could be held liable for an AI Overview that is objectively false. But it did not consider the summary at issue in that case to be false, and dismissed the application.

In another case, in Berlin, a court viewed AI Overviews not as independent content but as a new search format. In other words, courts have reached different conclusions about the nature of AI summaries. How far a feature that processes and presents search results can be treated as Google’s own speech is what divides the outcomes from case to case.

Does this matter for users in Korea?

Users in Korea have access to the same feature. Google announced that it would offer Korean-language AI Overviews in Korea starting December 5, 2024. The scenario at issue in the German cases is one where the summary that appears in search mischaracterizes a particular company or person. That affects not only the person searching, but also the companies and individuals named in the summary.

The core of these rulings is whether the operator of a service must answer for the content of AI-generated search summaries. The Munich court put weight on the fact that the AI evaluated information directly and rewrote it into sentences. What remains to be seen is the specific content of the Aachen decision, where the Munich case stands after the settlement, and how the conflicting court rulings will be reconciled.