A German court has ruled that Google is liable for false statements made by artificial intelligence reviews

Local court A company in Germany has issued a ruling that could reshape the work of search engines and AI-based chatbots around the world. The Munich Regional Court has tentatively ruled that Google is liable for a series of false statements generated by its AI Overviews feature, requiring the company to prevent false or inaccurate claims from being published through its search engine.

The ruling stems from a case first reported by Decoder, in which two publishers discovered that AI-generated Google summaries linked them, in some searches, to questionable business practices, scams, and subscription scams, without any basis for doing so.

Earlier this year, the affected companies sent a cease-and-desist letter to the tech giant, according to the report. Google denied responsibility, arguing that its auto-summary feature warns users that information may contain errors and should be independently verified.

The court’s analysis concluded that Google’s AI combined information corresponding to other companies flagged for potentially illegal practices with data from the plaintiffs, creating links that did not appear in any of the sources linked to the search engine.

The authorities found that Google’s tool, unlike traditional search engines, which only display lists of links to data provided by third parties, produced “independent, new and substantial data” based on a misinterpretation of information available on the Internet.

According to the court, it is not the responsibility of third parties to correct incorrect information. Google is the only entity with the ability to modify the technology underpinning its AI-generated summaries and, therefore, “must be held accountable.” Furthermore, the court found that Google’s line of defense lacked merit, because the challenged summary “contains data that does not appear at all in search results.”

A new (and powerful) interpretation of AI on the web

The court’s interpretation of the role of artificial intelligence in presenting research results could make this case a historic precedent. A large technology company has been found responsible for the impact of its most advanced developments on widely used platforms.

Until now, in most legal systems, search engines have been tools that merely facilitate access to content created by third parties and available on the web. This status gave them a certain level of protection when the information published was false, inaccurate, misleading or even defamatory.

However, the German court held that this guarantee no longer applies when search engines include generative AI systems. By their logic, this technology is capable of producing non-existent claims based on multiple sources, and therefore, the companies responsible for operating it must bear responsibility for the resulting content.

The judges also concluded that while Google encourages users to verify information because of the potential for hallucinations inherent in its AI models, this warning does not absolve the content distributor of liability. They believed that otherwise the victims of false statements would be defenseless, because the original sources never made the statements, and therefore could not be subjected to legal action.

Likewise, the court held that the results generated by an artificial intelligence system cannot be protected under the principles of freedom of expression, because they are the product of an algorithm designed, trained and managed by a company, and not the expression of an individual opinion.

As a precaution to prevent this from happening again, the ruling required Google to remove a significant portion of the data deemed defamatory in this case, and to cover 80 percent of the legal costs arising from the proceedings.

A company spokesperson, quoted by Ars Technica, suggested the decision could be appealed. “We invest deeply in the quality of our AI overview to ensure that the vast majority of responses provide accurate information and are designed to reflect information found on the web,” the statement read. He added: “We are carefully reviewing this decision, and it is not yet final.”