German Court Makes Google Liable for AI Answers

Google's AI Overviews falsely claimed that two Munich-based publishers were involved in scams, subscription traps, and "dubious business practices". The AI hallucinated information and made up connections that weren't in any of the linked sources. The publishers were not satisfied with Google's response to their cease-and-desist letter so they sued.

The verdict from the Regional Court of Munich was a temporary injunction barring Google from repeating those false claims and 80% coverage of the legal costs.

This is different because German search engines just point to third-party pages, but AI Overviews generate "independent, new, and substantive statements". These are Google's own words. The court called the false claims "the defendant's own statements" and said Google "alone has influence over the AI's offering and the algorithms with which the AI operates". So in plain English, Google's "AI can make mistakes" disclaimer doesn't work in this situation.

Google argued that users could check the linked sources and know not to trust AI blindly but the court said nope, and that the chance to fact-check a statement does not exempt whoever published it from liability. Studies show barely 1% of users click source links from AI Overviews anyway.

Google must remove the false content and ensure their AI doesn't generate false claims again. That should be easy. Also, this is a preliminary injunction from a regional court, not a final judgment, so Google can appeal.

The Problem Is Big

As per The New York Times, Google's AI Overviews are accurate ~91% of the time, but… more than half of even the correct answers were not supported by the cited sources. With about 2 billion people interacting with AI Overviews each month, the wrong ones add up to millions of wrong answers. That's a lot of chances to get sued.

This could happen with any other AI if/when similar cases are brought against them.