German court rules Google liable for AI-generated search summaries

A German court ruled this month that Google is liable for false statements in its AI-generated search summaries, rejecting arguments that users should verify AI outputs or that AI outputs are inherently untrustworthy. The court held that AI summaries are expressions of the company’s business activities, not neutral transmissions of third-party content.

The decision contrasts with traditional search, where courts have treated platforms as carriers facilitating access to third-party editorial content. Unlike search results that quote or link to sources, AI overviews rewrite and synthesize information, exercising editorial discretion similar to a publisher. This distinction was central to the ruling, which emphasized that AI summaries are not merely neutral tools but active representations of the company.

The ruling follows a 2024 case in which Air Canada was held liable for an AI chatbot’s incorrect promise of a discount, with the court rejecting the airline’s argument that the chatbot was a separate legal entity. Legal experts argue this sets a precedent that corporations are responsible for the statements made by their AI agents, whether human or automated.

The decision could force companies to either invest heavily in improving AI reliability or abandon high-risk use cases where errors could lead to legal liability. For example, AI-powered legal advisors, medical consultants, or media influencers may become commercially unviable if companies are held accountable for their outputs. The court’s reasoning implies that businesses cannot disavow responsibility for AI-generated statements simply because they are produced by software.

The ruling also raises broader questions about the applicability of Section 230 of the 1996 Communications Decency Act, which shields internet providers from liability for third-party speech. Courts have previously grappled with whether algorithmic curation of content transforms platforms from carriers into publishers. The German decision suggests that AI-generated content may not fit neatly into existing legal categories, requiring new frameworks for accountability.