Who is liable,

if the

Google AI lies.

23. July 2026

of

Clemens Pfitzer

Can Google be held legally liable if its AI spreads false claims and lies about your business? How do you defend yourself if Google’s “AI Overview” implies that your business is involved in fraud?

What happens when Google does more than just search?

For some time now, when you search for a company on Google, you’ve been getting not just a list of links—but sometimes a ready-made answer as well. Above the actual search results, labeled “AI Overview,” a summary appears that directly answers the search query. Sounds convenient. But it can have devastating consequences for the company in question—especially if the text is simply wrong.

That’s exactly what happened to a German publishing house. Following a search query that combined the company’s name with the term “scam,” the AI overview displayed a statement claiming that the company was “known for unscrupulous business practices” and was “often perceived as a scam.” Specific allegations were also listed: subscription traps, fake phone calls, failure to unlock paid content, and frequent name and URL changes. None of this had been substantiated.

What the Court Ruled

The Munich I Regional Court subsequently ordered Google to Judgment of May 28, 2026 – Case No. 26 O 869/26 ordered to cease and desist by way of a preliminary injunction. The court has prohibited Google from disseminating the disputed statements about the plaintiff companies—including allegations of scams, subscription traps, locked-behind-paywall content, poor accessibility, and alleged connections to other companies. However, the motion was partially dismissed, for example, with regard to the alleged collaboration with a debt collection agency—in this instance, the publishing house had not sufficiently demonstrated that the claim was false.

The court based the plaintiffs’ central claim on the right to corporate personality. It viewed the AI-generated statements as an unlawful infringement on the companies’ reputation in business dealings, which could not be justified by any legitimate interest on Google’s part.

In doing so, the court made it clear that the AI overview is of a different nature than a conventional search result.

The Difference Between a Search Engine and an AI Overview

Until now, search engines have been subject to a relatively lenient standard of liability. Anyone who merely displays links to third-party content is generally not responsible for what appears on the linked pages—and is only required to take action once explicitly notified of an obvious legal violation. For a long time, this protection served as the legal basis on which Google could rely in disputes over search results.

The court has now denied this protection for the “AI summary.” The key difference: The AI summary is not merely a reproduction of third-party content, but an independent expression. The AI evaluates, weighs, and structures the information—and in doing so arrives at statements that do not appear at all in the linked sources. In this specific case, for example, the AI response contained claims about connections to other companies that were not found in any of the displayed links. The court therefore concluded that Google must be held responsible for these statements as its own.

The court also rejected Google’s argument that users could verify for themselves, using the linked sources, whether the AI summary was accurate. It pointed out that the “AI summary” presents a self-contained, self-explanatory statement—without any indication of possible errors or uncertainties. Someone who reads a newspaper article based solely on the headline shouldn’t have to be told that they should have kept reading. The same applies to the AI summary.

The court also pointed out that the AI feature—unlike the search engine itself—is not strictly necessary for using the Internet. A search engine that merely displays links is already useful enough to make the flood of information manageable. The AI summary is an additional feature—which Google can therefore operate even without the special liability protection that was developed for search engines as socially indispensable infrastructure.

Finally, the court also rejected the argument that the provisions of the Digital Services Act protected Google from liability. Insofar as the statements in question are the provider’s own attributable statements, the liability exemption for hosting providers does not apply. Civil law injunctive relief claims remain expressly unaffected.

Why the statements were unlawful in this specific case

The court carefully distinguished between the contested statements based on their nature: Some were classified as statements of fact, while others were classified as expressions of opinion.

With regard to factual claims, the following applies: False claims that could damage a company’s reputation in business dealings are generally unlawful. The decisive factor here was that Google could not prove that the alleged connections to other companies or the practices described were in fact true. The company had affirmed under oath that it had no connection to the companies in question. Google had not presented any substantial counterarguments.

Even in the case of statements that the court classified as expressions of opinion, a claim for an injunction was granted. While expressions of opinion enjoy significant constitutional protection, that protection ends where the opinion is based on false facts. Anyone who calls someone a fraud because they falsely associate that person with certain companies cannot invoke freedom of speech. The court also emphasized that AI-generated statements should not be regarded as expressions of personal conviction anyway, but rather as the result of an algorithm—and thus carry correspondingly less weight in terms of freedom of expression.

Conclusion

We are happy to

advise you about

AI!

Contact form

Call now!