Did AI Platforms Just Lose Their Liability Shield?
On June 9, a German court ruling found that Google’s AI Overviews are Google’s own speech, not neutral search results, and held the company legally responsible for the false statements that its AI generated. For boards and senior executives, the ruling reframes AI visibility as a governance and legal concern, not just a marketing one.
Listen to “Did AI Platforms Just Lose Their Liability Shield?” Deep Dive
Last week, a German court issued a preliminary injunction against Google that found that Google’s AI Overviews are Google’s own words, making the company liable for false answers generated by the platform. This is despite the fact that large language models (LLMs)—e.g., Google Gemini, ChatGPT, Claude and others—provide a disclaimer below their search field warning users that the platform can make mistakes and that they should verify important information. Until now, that disclaimer was meant to shield platforms like Google, OpenAI and Anthropic from rulings like this one.
The dispute arose because Google’s AI had invented connections that appeared in none of the linked sources, mixing the plaintiffs (two publishers) up with genuinely shady firms. The publishers sent Google a cease-and-desist letter, but Google did not respond adequately.
“If AI platforms can be held liable for what their AI says about a company, then what AI says about your company is no longer just a brand issue; rather, it is one that companies and boards now have a real reason to monitor.”
The biggest surprise here is how the German court legally reclassified Google’s AI as a publisher rather than as a search engine. Search engines have long enjoyed limited liability because they merely point users to third-party pages. AI Overviews, the court found, do something fundamentally different: They generate independent, new and substantive statements in Google’s own voice. The platform alone controls what those statements say, and the court ruled that the platform is therefore the author of those statements, rather than a neutral intermediary.
The court also rejected Google’s central defense, which was that users can check the linked sources themselves and know not to trust AI blindly. The ability to disprove a statement through further research, the court said, does not exempt whoever published it from liability. The court drew a parallel to press law, where a misleading teaser is actionable even if no reader clicks through to the full article.
For boards and senior executives, this ruling adds a new layer of AI governance complexity. Until now, AI visibility has been treated as a marketing and competitive concern. The Munich ruling reframes it as a legal one. If AI platforms can be held liable for what their AI says about a company, then what AI says about your company is no longer just a brand issue; rather, it is one that companies and boards now have a real reason to monitor.
How Does This Reclassification Affect Your Company?
The Munich ruling has implications well beyond Google. Any company building or deploying its own AI tool faces the same legal exposure. Think of Amazon’s Rufus, Walmart’s Sparky, customer service chatbots and internal search agents. Every one of these tools generates statements that, under the Munich court’s logic, could be considered the company’s own speech rather than neutral search results.
For boards, the question shifts from “this doesn’t concern us” to “where is our AI tool pulling answers from, and is that making us legally vulnerable?”
The Scale of the Exposure Is Significant
A New York Times analysis found that Google’s AI Overviews, running on Gemini 3, were accurate 91% of the time. That sounds reassuring until you look at the other half of the finding: More than half of even the correct answers were not supported by the sources cited. At Google’s scale, the 9% error rate translates into millions of potentially false or unsupported answers. Each one is now potentially a legal liability rather than a search result.
For a company deploying its own AI, the math is the same. Even a 1% error rate at customer-service scale produces hundreds of false statements per day. Each one is now potentially considered to be the company’s own speech.
The disclaimer that “AI can make mistakes” no longer offers the protection it once did. Boards and senior executives need to pivot their AI strategies from treating AI as a passive resource to treating it as an editor whose job is to scrutinize, question and validate the accuracy of what the AI says.
Three Questions Boards Should Be Asking Now
The Munich ruling doesn’t require boards to become AI experts; rather, it requires them to ask better questions about something they’re already supposed to govern. Here are three questions that are worth putting on the agenda.
1. What is AI currently saying about our company, and is it accurate?
Before the Munich ruling, this was a marketing concern, which has now transitioned into a governance concern. If AI platforms are publishing statements about your company that are false, misleading or damaging, and you have no process in place for detecting those statements, you have no way to dispute them and no record of having tried. That’s both a legal issue and a reputational one.
2. Where does our own AI pull its answers from, and who reviews its outputs?
If your company has built or is building an AI tool, such as a customer service chatbot or a recommendation engine, the Munich logic applies to your AI the same way it applied to Google’s. Boards should know what data the AI was trained on, what sources it’s drawing from and who reviews its outputs before they reach customers or partners. The “AI can make mistakes” disclaimer is no longer enough.
3. How quickly can we respond when an AI says something false about our company or through our company’s tools?
If a journalist, customer or competitor flags an AI-generated falsehood about your company, what’s the process for documenting it, disputing it with the platform and tracking the response? If the answer is “we don’t have one,” that’s a gap that needs to be closed before the next ruling lands.
At Your Next Board Meeting…
Ask the person leading your AI initiative this question: “What’s our process for verifying what AI platforms say about us, and how quickly can we respond when an inaccuracy is flagged?”
The answer will tell you everything about whether your AI governance is keeping up with how AI law is changing.
If you’re not sure what AI is currently saying about your company, that’s the place to start. I’m running a small number of AI visibility diagnostics through introductions from board directors and senior advisors. If that’s a conversation worth having for your company, email me at mike@scheinerinc.com.
As always, thank you for reading and/or listening.
