German courts rule Google liable for what its AI search says

A German regional court in Aachen has ordered Google to stop repeating AI-generated claims that a prominent neuroradiology professor was suspended and convicted of sexual coercion - claims the court found referred to a different person. Google’s AI Mode produced a detailed, false account; deletion and a legal warning were ignored, so the court issued an injunction without hearing Google because the allegations were publicly visible to an unlimited audience. Google Ireland faces fines up to €250,000 per breach or potential detention of managers if it republishes three variants of the text. The court held that Google “provides AI-generated search and is therefore also responsible for the answers generated,” noting the company could have checked and corrected the AI’s output once alerted.

The ruling follows a similar Munich decision that treated Google’s AI Overviews as independent, substantive statements for which Google must answer, a case later settled on appeal so it is not binding precedent. By contrast, a Berlin court ruled in a trademark dispute that AI summaries are extended search results and Google remains a neutral intermediary; U.S. courts have also diverged, dismissing related publisher claims on other grounds. The practical consequence is clear: if AI answers count as Google’s own statements, the company can carry publisher-like liability for false claims and refusals to take down defamatory output.