German Ruling Declares Google Liable for AI Overviews' Wrong Answers

Munich Regional Court in Germany ruled that Google must bear direct legal liability for false information generated by its AI Overviews. The court determined that unlike traditional search engines, which simply link to external websites, AI summaries reconstruct information to generate their own content with independent sentences and structure.

Key points of this ruling include the following:

• Differentiation from search results: Because AI summaries evaluate and combine third-party content to generate new statements, the 'limited liability doctrine' applied to traditional search engines cannot be applied.
• Limits of user verification responsibility: Google argued that users can directly check the sources, but the court found that since AI summaries themselves contain self-contained statements, liability is not exempted merely by the possibility of further investigation by users.
• Expanding legal risk: Even if the accuracy of the Gemini 3-based model reaches 91%, given Google's scale, the millions of wrong answers occurring every hour could spread legal risk across similar AI services such as ChatGPT, Claude, and Perplexity.

This ruling holds significant meaning in that it explicitly states that when AI-generated information infringes on the personal rights of a plaintiff (publisher), the platform operator must bear responsibility not as a mere hosting provider but as a content creation subject.