A German court has issued a significant ruling on the legal responsibility of artificial intelligence (AI) systems, holding that Google can be held accountable for false or defamatory information generated by its AI Overviews feature.
The case, reported by The Decoder, was brought by two German publishers who alleged that Google’s AI-generated summaries falsely described their businesses as scams and accused them of engaging in questionable practices.
In its ruling, the court drew a distinction between traditional search engines and AI-generated responses.
Judges noted that conventional search engines primarily direct users to third-party websites by displaying links and snippets of content created by others.
AI Overviews, however, generate original responses based on information gathered from multiple sources.
According to the court, AI-generated summaries are not merely reproducing existing content but are creating “independent, new and substantive statements.” As a result, the court concluded that AI-generated outputs should be treated differently from traditional search results.
The judges stated that only Google has the ability to modify, correct, or improve the underlying AI system that produces these responses.
Therefore, the court found that inaccurate or defamatory statements generated by AI Overviews form part of Google’s own commercial activity, making the company potentially liable for those outputs.
The court also rejected Google’s argument that users should independently verify AI-generated information before relying on it. The judges observed that requiring users to fact-check every response would significantly reduce the usefulness of AI-generated summaries.
Google argued that users generally understand that AI systems are not always accurate and that AI-generated content may contain errors.
The company pointed to warning labels and disclaimers used across the AI industry, which advise users to verify information independently.
Following the ruling, Google stated that it invests heavily in improving the quality and accuracy of AI Overviews and that the vast majority of responses provide accurate information. The company also noted that the decision is not final and remains subject to further legal review.
The case has drawn international attention because it addresses a broader legal question surrounding generative AI systems.
Unlike traditional platforms that host or display content created by others, generative AI tools such as Google’s AI Overviews, OpenAI’s ChatGPT, Microsoft’s Copilot, and Anthropic’s Claude generate original responses based on information available to them.
The German ruling is one of the clearest judicial decisions so far addressing whether companies operating AI systems can be held responsible for inaccurate information generated by those systems.
While the decision currently applies only within Germany, it is expected to contribute to ongoing discussions about AI regulation and liability in multiple jurisdictions around the world.
