Court holds Google liable for AI Overviews
A Munich court ruled Google is directly liable when its AI Overviews invent false claims, treating the summaries as its own words. Google is appealing.
The Regional Court of Munich has ruled that Google can be held directly liable when its AI Overviews state things that are not true. The May 28 decision (case 26 O 869/26) sided with two Munich publishers whose businesses the AI summary had tied to scams, subscription traps, and dubious practices, accusations that appeared nowhere in the underlying search results.
The reasoning is the part that travels. The court declined to treat the AI Overview as a neutral list of links and instead called it Google’s own speech, an independent and substantive statement the company is answerable for. That turns a hallucination from someone else’s words misquoted into Google’s own defamation. Google says its overviews are built to reflect what is already on the web, that it invests in quality, and that the judgment is not final. It is appealing.
Why this matters
The liability question many operators have waved away just got a concrete answer in one courtroom: if an AI feature publishes in your name, its mistakes can become your mistakes. The ruling is German and under appeal, so read it as a direction of travel, one more sign of the legal scrutiny circling AI, from state attorneys general to the courts. The cheap precaution is the one we keep returning to in the verification habit: treat model output you put in front of customers as something you authored. If automated summaries or support replies speak for you, the lesson from what AI gets wrong is to keep a person between the model and anything that names a real customer, vendor, or competitor.