German court says Google is liable for false AI Overviews statements

In short: A court in Munich, Germany, has preliminarily ruled that Google is responsible for false statements produced by its AI Overviews in Search.

What happened

The Munich Regional Court looked at a case involving two publishers who said Google’s AI Overviews wrongly linked them to scams and subscription fraud. According to the reports, those accusations had no basis in the websites that Google linked to.

Google argued it should not be liable because AI Overviews come with a warning that the information may contain errors and should be checked. The court did not accept that defense. It said the summary included statements that did not appear in the search results at all.

The court also said AI Overviews are not just a list of links like traditional search. Instead, the tool creates new sentences by mixing information from different places. The court found that it combined information about other companies that had been flagged for suspicious behavior with data about the publishers, and produced a misleading association (like putting the wrong label on the wrong package).

As a precaution, the court required Google to remove many of the statements it viewed as defamatory, and to pay 80 percent of the legal costs. A Google spokesperson said the company is reviewing the decision, and noted the ruling is not final and could be appealed.

Why it matters

Many AI tools tell users, “This might be wrong.” This ruling says a warning may not be enough when a company’s system generates claims that were not in the original sources. If this approach spreads, it could change how search engines and AI chatbots handle mistakes, and it could make it easier for people and businesses to challenge harmful false information.

Source: Wired