German Court Rules Google Can Be Held Liable for AI-Generated Misinformation
The Munich Regional Court has preliminarily ruled that Google may be held liable for false and defamatory statements generated by its AI Overviews feature after the system incorrectly linked two publishers to fraudulent business practices and scams. The court found that the AI-generated claims did not appear in the original search results and were instead created by Google's technology through the misinterpretation and combination of information from multiple sources.
The decision marks an important shift in how courts may view AI-generated content. Traditionally, search engines have been treated as intermediaries that simply provide access to information published by third parties. However, the court concluded that generative AI systems go beyond displaying existing content by creating new statements and interpretations, making the operators of those systems responsible for the resulting output.
Google argued that users are warned AI-generated summaries may contain errors and should be independently verified. The court rejected this defence, finding that such disclaimers do not remove responsibility when the technology generates inaccurate or defamatory information.
The ruling required Google to remove the disputed content and cover a significant portion of the legal costs associated with the case. While the decision may still be appealed, it has already sparked discussion across the technology and legal sectors.
For records and information management professionals, the case highlights growing questions about accountability, evidentiary integrity and governance in AI-generated content. As organisations increasingly rely on AI tools to create, summarise and interpret information, ensuring the accuracy, traceability and reliability of information sources will become increasingly important.
The ruling may also influence how other AI providers approach governance and risk management. If upheld, it could establish a precedent that organisations developing, training and operating AI systems may be legally responsible for the content their technologies generate, even when those outputs are accompanied by warnings or disclaimers.
As AI adoption continues to accelerate, the case serves as a reminder that trust in information remains a fundamental governance issue and one that organisations cannot afford to overlook.