Recently, a German court ruling caught my attention because it asserts that Google can be directly liable for false claims made in their AI Overviews. The Regional Court of Munich’s decision highlights a significant shift, considering AI-generated summaries as Google’s own content rather than just protected search results.
This ruling emerged from a case where AI Overviews mistakenly linked two Munich publishers to scams and dubious practices, despite the linked pages containing no such evidence, as reported by The Decoder.
AI Overviews are not just search tools. According to the court, these Overviews go beyond merely assisting users in finding third-party content. They actually process and present information in their own distinctive manner.
What struck me was the court’s findings that the AI Overview allegedly made standalone accusations regarding questionable business practices, which were not substantiated by the linked sources. Because Google crafts and controls these features and their algorithms, the court ruled these statements to be Google’s own content.
Traditional search protections didn’t apply here. Google argued that they should be protected by German case law, which generally shields search engines as indirect infringers. However, the court disagreed, emphasizing that AI Overviews are distinct as they generate new statements from multiple sources.
The court also dismissed Google’s argument that users could verify claims by reviewing linked content. They highlighted that AI Overviews offer claims that stand as complete answers without needing verification.
Why does this matter to me? The court’s stance implies that AI Overviews aren’t neutral links. If they issue incorrect claims about a company, Google may bear direct responsibility for these words.
Mismatched connections and misinformation. The court determined that misinformation resulted from AI conflating data about other entities with that concerning the publishers.
Given that the contested claims weren’t present on the linked sites, the publishers lacked a clear third party to target legally, should Google be considered only as an intermediary.
Interestingly, the court insisted that Google could compare AI-generated content against primary sources, at least in analogous situations.
Action required from Google. The injunction demands that Google refrains from repeating the disputed claims, which include allegations of scams and nonexistent business practices.
Furthermore, Google is instructed to bear 80% of the legal costs, while each publisher covers 10%. Despite Google’s lack of a cease-and-desist declaration with a penalty clause, the potential for repeat violations was noted, emphasizing the importance of this ruling for future similar claims.
Inspired by this post on Search Engine Land.
