A German court draws a narrower line around AI search liability
A Berlin court has ruled that Google’s AI-generated Overviews should be treated as a new search-result format rather than as original statements made by Google itself. The decision, described by The Decoder, says users can recognize that the system is compiling information from third-party sources and that Google has no decisive influence over the content in the way the plaintiff alleged.
The case arose from a lawsuit filed by a perfume company over trademark issues. According to the supplied text, Google’s AI summary mentioned protected brand names alongside cheaper imitation products and linked to websites selling those alternatives. The Berlin court dismissed the claim, treating the output as an aggregation of information already available elsewhere on the web.
Why the ruling matters
The importance of the decision is less about perfume than about a core question facing generative search: when does an AI summary become the platform’s own legally attributable speech? Berlin’s answer is comparatively restrictive. If users would understand the result as a synthesized presentation of outside sources, then Google’s role looks closer to display and organization than authorship.
That is a consequential position because it could limit platform liability in at least some classes of disputes. If adopted more broadly, it would support the view that AI Overviews are an interface layer on top of indexed web content, not an independent publisher standing behind every claim in the generated response.
Berlin and Munich are now in tension
The ruling also stands in clear contrast with a recent Munich decision summarized in the same source text. In that case, a court reportedly treated Google’s AI summaries as independent content after the system produced false factual claims linking publishers to fraudulent schemes. Munich held Google liable because only Google controls the AI system and its algorithms.
Taken together, the two cases show that the legal framework for generative search remains unsettled. One court sees summarized AI output as a new search format. Another sees it as independently actionable content when the system makes claims not found in the cited sources. That split matters for product design, disclosure, moderation, and litigation strategy across the sector.
A foundational question is still open
The broader issue is not unique to Google. Any company building generative search tools will face similar arguments about authorship, control, and user expectations. Courts will need to decide whether synthesis changes the legal character of information display, and if so, under what conditions.
The Berlin ruling does not resolve that debate. It makes it more visible. For now, the practical lesson is that AI search liability in Europe is still taking shape case by case, with different courts applying different intuitions about how much agency these systems really have.
Why the decision is significant
- Berlin treated AI Overviews as a search display format, not original content
- The case involved trademark complaints over perfume imitations
- A recent Munich ruling reportedly reached the opposite conclusion in a false-claims case
- The split leaves generative-search liability legally unsettled
This article is based on reporting by The Decoder. Read the original article.
Originally published on the-decoder.com
