Munich Regional Court issued an injunction holding Google liable for defamatory AI-generated summaries.
Why it matters: Legal professionals must note this ruling expands liability for AI-generated content beyond traditional search results protections. It clarifies responsibilities for corporations deploying generative AI tools under European law.
- On May 28, 2026, Munich Regional Court granted an injunction against Google over defamatory AI-generated summaries called AI Overviews.
- The court ruled the AI Overviews are Google's own content, removing usual search engine liability shields.
- Two Munich publishers were falsely linked to scams with no evidence in cited sources.
- Google must cover 80% of legal costs after ignoring cease-and-desist demands from the publishers.
On May 28, 2026, the Munich Regional Court held Google liable for defamatory statements produced by its AI feature known as AI Overviews. These automatically generated summaries falsely linked two Munich-based publishing houses to scams and illicit business practices, despite the web sources cited not supporting these claims.
The court distinguished AI Overviews from typical search engine results by classifying them as Google's own content, rejecting the notion that the summaries are mere third-party information. This removes Google's usual protections under the German Telemedia Act, which typically shields search engines from liability for third-party content.
AI Overviews are generated algorithmically via Google's Gemini 3 language model, which independent tests have identified as having approximately a 9% factual error rate and wrongly attributing sources about 56% of the time, as reported by the AI Accuracy Study 2026. These inaccuracies pose significant reputational risks and legal liabilities.
Before the court ruling, affected publishers sent Google cease-and-desist letters requesting corrections, which Google did not act upon. Consequently, the court ordered Google to pay 80% of the legal costs. A Google press release stated the company is reviewing the judgment and is committed to enhancing AI content accuracy.
This decision marks a significant precedent in European AI regulation and liability, signaling heightened scrutiny for companies deploying generative AI that produces public-facing content. Legal counsel and compliance teams at corporations and law firms should reassess risk management, content oversight, and potential liability associated with AI-generated summaries and information services.
By the numbers:
- 9% — estimated factual error rate of Google's Gemini 3 language model, per AI Accuracy Study 2026
- 56% — rate of incorrect source attributions in AI Overviews, per AI Accuracy Study 2026
- 80% — portion of legal costs Google was ordered to pay following injunction
Yes, but: While the ruling imposes liability for AI-generated content, Google is appealing the decision, and the scope of similar regulatory actions across other EU jurisdictions remains uncertain.
What's next: Await developments from Google's appeal process and monitor similar cases as European regulators increasingly focus on generative AI accountability.