In a preliminary ruling with potentially far-reaching consequences for the AI industry, a German court has determined that Google can be held legally responsible for defamatory or inaccurate content generated by its AI Overviews feature.
The decision marks what appears to be one of the first times a court has held an AI company directly accountable for the “speech” produced by its systems, challenging long-held assumptions about platform liability in the age of generative tools.
The case, brought by two German publishers, centered on AI Overviews that falsely linked them to scams and dubious business practices. Google’s system generated statements such as “Yes, [it] is known for dubious business practices and is often perceived as a scam,” even though those claims did not appear in the underlying search results. After the publishers sent a cease-and-desist letter earlier this year, Google failed to correct the misleading outputs, prompting the legal action.
Register for Tekedia Mini-MBA edition 20 (June 8 – Sept 5, 2026).
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab.
The court rejected Google’s defense that users understand AI outputs are not always accurate and should be verified. It emphasized a key distinction: while traditional search engines merely surface links to third-party content, AI Overviews create “independent, new, and substantive statements” based on the system’s own interpretation of web data. Because only Google controls the underlying algorithm and can correct erroneous outputs, the company bears responsibility when it fails to act.
The ruling requires Google to stop displaying the false claims in AI Overviews, at least temporarily. It also noted that the tool’s utility would be undermined if users were forced to independently verify every link, directly countering Google’s argument that disclaimers are sufficient protection.
The decision arrives amid growing evidence of reliability issues with AI-powered search. A May analysis by The New York Times found that Google’s AI Overviews using the current Gemini 3 model were inaccurate about 9% of the time and included inaccurate source links about 56% of the time. A Pew survey from last July showed that most users do not click through to verify the sources cited in AI summaries, increasing the potential for misinformation to spread unchecked.
These findings suggest that AI Overviews could be generating millions of flawed or misleading responses daily, amplifying risks for individuals, businesses, and public discourse. The German court appeared to recognize this danger, ruling that publishers’ interest in correcting false information outweighed Google’s commercial speech rights in this context.
The ruling could set a precedent that influences how courts worldwide view liability for generative AI. For years, AI companies have relied on disclaimers warning users about potential inaccuracies, hoping to shield themselves from legal responsibility. Some have even argued that AI-generated content constitutes a new category of “pure speech” deserving strong First Amendment-like protections.
The German decision pushes back against that position. It treats the false outputs as “primarily an expression of the defendant’s commercial activity,” making Google accountable for the content it actively produces and presents. This reasoning could resonate in other jurisdictions, particularly in Europe, where data protection and digital services regulations already impose stricter obligations on tech platforms.
For Google and other AI search providers, the case highlights a fundamental vulnerability that, unlike traditional search, AI summaries are not strictly necessary for users to find information online. The court noted that people can still navigate the web effectively without AI Overviews, weakening arguments that such tools deserve broad immunity because they help manage an overwhelming “flood of data.”
A Google spokesperson told Ars Technica that the company is “carefully reviewing this decision, which is not yet final.”
“We invest deeply in the quality of AI Overviews to ensure that the overwhelming majority of responses provide accurate information, and they are designed to reflect the information that exists on the web,” the spokesperson said, emphasizing ongoing investments in quality control.
Google is also working with the FBI and telecom partners AT&T, T-Mobile, and Verizon to dismantle related phishing infrastructure in a separate but connected effort against cybercrime. The company is endorsing seven pending U.S. congressional bills aimed at countering online scams.
Despite these steps, the German ruling could encourage more publishers, individuals, and organizations to pursue legal action when AI tools generate harmful or false content. Some analysts believe that if similar decisions follow in other countries, major AI developers could face a wave of lawsuits during what many see as an experimental phase of AI search deployment.