A Munich court held that Google must answer for the statements in its AI‑generated search overviews, rejecting the shield that covers ordinary search engines. The ruling bars the company from repeating false claims that tie local publishers to scams, calling them a direct infringer because the summaries are Google’s own content.

Under German law, AI overviews create independent, substantive statements that Google can control, unlike a list of links. The court noted that users rarely click the cited sources and that the AI’s claims were not found in any linked page, making them Google’s own words. This distinguishes AI content from traditional search results.

The decision forces Google to monitor its Gemini‑based summaries and may prompt other AI providers to tighten source verification. With a 91 percent accuracy rate, the court warns that millions of wrong answers can still surface, potentially sparking defamation claims. The ruling signals that AI‑generated opinions receive less free‑speech protection than plain search listings.

Google paid 80 percent of the legal costs, while the publishers absorbed a tenth each, reflecting the court’s view that the company bears primary responsibility. The injunction remains temporary, but the case could influence international standards for AI liability, especially as other platforms like ChatGPT and Claude adopt similar summarization features.