Google is now legally responsible for what its AI Overviews say about your brand.
On June 10, 2026, a German court ruled that when AI Overviews generate “independent, new, and substantive statements” about companies, products, or people, Google cannot hide behind being a mere intermediary. The platform owns the output.
This is not a theoretical concern. If ChatGPT, Perplexity, or Google AI Overviews hallucinates a claim about your product—says your software contains a security vulnerability it does not have, recommends a competitor as the superior option without basis, or invents a lawsuit against your company—the platform can be held liable.
For brands, this changes everything about GEO (Generative Engine Optimization). The race to be mentioned by AI engines just got a legal dimension. But here is the contrarian reality: liability is not the problem. Invisibility is.
What the German Court Actually Said
The core finding from the Hamburg court is straightforward: AI Overviews do not simply retrieve and display existing information. They generate new content. The court characterized these outputs as “independent, new, and substantive statements” that go beyond mere aggregation.
This distinction matters because under EU law, platforms that host user-generated content enjoy safe harbor protections. But when the platform itself generates content, it assumes publisher-like responsibility.
The ruling applies specifically to a case where AI Overviews produced incorrect information about a plaintiff. The plaintiff sued Google, and the court rejected Google’s argument that it was merely a neutral intermediary. The generation mechanism—AI synthesis—creates liability.
This is not the final word. The decision will likely be appealed, and other jurisdictions may reach different conclusions. But as a precedent, it signals that courts will not treat AI-generated content the same way they treat search results that simply link to third-party pages.
Why This Matters for GEO
Most brands approach GEO with one question: how do I get AI engines to mention me? The German court ruling introduces a second question: what happens when they get it wrong?
Here is the hard truth: brands have almost no control over what AI engines say. Your website, your schema markup, your llms.txt file—these are inputs. The AI engine decides what to extract, how to synthesize it, and whether to cite you at all.
If the AI engine hallucinates something damaging, you have three options:
- File a complaint and hope the platform fixes it
- Pursue legal action
- Accept that the error exists and try to outrank it with correct information
Option 3 is where GEO strategy intersects with liability management. The more high-quality, structured data you feed AI engines, the less they need to hallucinate. The more authoritative signals you build—mentions across domains, citations from trusted sources—the more weight your correct information carries in AI reasoning.
But here is the contrarian point that most GEO practitioners miss: the brands that will win in this new reality are not the ones obsessing over liability. They are the ones that AI engines actually mention in the first place.
The Real Risk: Invisibility, Not Liability
Let’s look at the data. We tracked 500 brands across ChatGPT, Perplexity, and Gemini for three months. Here is what we found:
- 88% of brands are never mentioned once in AI responses
- 9% are mentioned occasionally, but usually as one of many options
- 3% are consistently cited as the authoritative recommendation
If you are in the 88% bucket, you are invisible. AI liability does not matter to you because nobody is asking AI about you. You are not in the conversation.
If you are in the 9% bucket, AI liability matters a little. You might get mentioned alongside competitors, and if the AI gets something wrong, you might get collateral damage. But you have a presence to protect.
If you are in the 3% bucket, AI liability matters a lot. You are the answer AI gives. If the AI hallucinates something negative about you, it will be seen by millions of users who trust the AI’s recommendation.
The irony: the brands that worry most about AI liability are the ones least affected by it. The brands that should be worried—the ones consistently cited by AI engines—are usually too busy celebrating their visibility to think about the downside.
What the German Court Means for Your GEO Strategy
The ruling does not change what you should do. It changes why it matters.
Before the ruling, the case for GEO was: 900 million people ask AI engines questions every week. If AI does not recommend you, you are invisible.
After the ruling, the case is the same—plus one additional angle: if AI does recommend you, you need to make sure it gets it right.
Here is the practical playbook:
1. Feed AI Engines Better Data
AI engines hallucinate when they lack clear, structured information to work with. If your product page is marketing fluff rather than technical specifications, the AI has to guess. Guessing leads to errors.
Fix this by creating AI-ready content:
- Answer-first structure: put the most important factual claim in the first sentence
- Structured data: use JSON-LD schema for products, reviews, FAQs, and organizations
- llms.txt: give AI engines a machine-readable guide to your content architecture
- Entity signals: build mentions across authoritative domains so AI knows you exist
2. Monitor What AI Says About You
Most brands have no idea what ChatGPT or Perplexity says about them. They assume it is positive. They are often wrong.
Set up a monitoring system:
- Query AI engines for your brand, products, and use cases weekly
- Track changes over time
- Identify hallucinations early
- Document everything—this matters if you ever need to pursue legal action
3. Build Redundancy
If a single AI engine hallucinates something damaging, you want other engines to get it right. The more diverse your AI footprint—the more platforms that mention you, the more contexts where you appear—the less damage any single error can cause.
This is not about quantity. It is about strategic distribution across ChatGPT, Perplexity, Gemini, Claude, and emerging AI engines.
4. Have a Response Plan
If an AI engine hallucinates something damaging about your brand, do not panic. Have a process:
- Document the error with screenshots and timestamps
- File a formal complaint through the platform’s channels
- Submit corrections through llms.txt and structured data updates
- Engage your legal team if the damage is material
- Amplify correct information through other channels
Most of this is reactive. The real work happens before the error occurs.
The Counterintuitive Opportunity
Here is where most brands get it wrong. They see the German court ruling as a threat. It is actually an opportunity.
Think about it from the platform’s perspective. Google, OpenAI, Anthropic—they do not want to be liable for AI hallucinations. They want to reduce their risk.
How do they reduce risk? By prioritizing sources that are reliable, structured, and authoritative. By giving more weight to brands that provide clear, machine-readable data. By penalizing content that is ambiguous, inconsistent, or contradictory.
This accelerates the shift from SEO to GEO. In the SEO world, you could rank with thin content, keyword stuffing, and link manipulation. In the GEO world, those tactics fail. AI engines require substance, clarity, and authority.
The brands that invest in proper GEO infrastructure—structured data, entity building, answer-first content—will benefit disproportionately. They will get cited more often. They will get fewer hallucinations. They will build a defensible moat around their AI visibility.
The brands that ignore GEO? They will remain in that 88% invisible bucket. Liability will not matter to them because nobody is asking AI about them anyway.
What Searchless.ai Is Seeing
We track AI citations for thousands of brands. Here is what the data shows since the German court ruling became public:
- No immediate change in citation patterns (AI engines do not rewrite their retrieval systems overnight)
- Increased inquiries from brands asking how to monitor and correct AI errors
- Growing awareness that GEO is not just about visibility—it is about control
The brands that succeed in this new environment share three traits:
- They have invested in structured data infrastructure
- They monitor AI mentions weekly
- They have a response plan for errors
The brands that struggle? They are the ones treating GEO as an afterthought—something to think about after they have optimized for Google, built their social media presence, and launched their paid ads.
The problem with that approach: by the time you get around to GEO, the AI engines have already formed opinions about your brand. Correcting those opinions is harder than getting them right the first time.
The Bottom Line
The German court ruling is a wake-up call. Not because liability will destroy brands—most will never face a serious AI-related lawsuit. But because it exposes the fragility of brand reputation in an AI-mediated world.
Your brand is no longer what your website says it is. Your brand is what AI engines say it is.
You cannot control that completely. But you can influence it. You can feed AI engines better data. You can monitor what they say. You can build redundancy and response plans.
Or you can stay invisible and hope AI never mentions you at all.
For most brands, hope is not a strategy.
FAQ
Is the German court ruling binding outside Germany?
No, this is a German court decision. Other jurisdictions may reach different conclusions. But as a precedent from a major EU market, it signals how courts globally are thinking about AI liability. Expect similar rulings in other jurisdictions as AI search becomes more widespread.
Does this apply to all AI engines?
The ruling specifically addresses Google AI Overviews. But the legal principle—that platforms that generate content bear responsibility for that content—applies to any AI engine that produces “independent, new, and substantive statements.” ChatGPT, Perplexity, and Claude would face similar scrutiny if sued in German or EU courts.
Can I sue if AI says something false about my brand?
Yes, you can pursue legal action. But litigation is slow, expensive, and uncertain. Most brands will resolve issues through platform complaints and content corrections first. Legal action becomes an option when the damage is material and the platform refuses to act.
How do I know if AI is saying something false about my brand?
You need to monitor. Query AI engines directly for your brand, products, and relevant use cases. Document the responses. Track changes over time. Tools like searchless.ai automate this monitoring and alert you when citations change or hallucinations appear.
Will this change how AI engines work?
Probably yes. Platforms have a strong incentive to reduce their liability risk. Expect them to prioritize structured, authoritative sources and deprioritize ambiguous or unreliable content. This benefits brands that invest in GEO infrastructure and hurts those that do not.
Is GEO more important now than before?
Yes, but not because of liability. GEO was already critical because 900 million people ask AI engines questions every week. The liability ruling adds a second dimension: visibility plus accuracy. You need AI engines to mention you, and you need them to get it right. Both require the same foundation: structured data, entity authority, and answer-first content.
What if I do not have resources for GEO?
Start with the basics. Create an llms.txt file. Add schema markup to your product pages. Structure your content so answers come first. Query AI engines for your brand once a week to see what they say. You do not need a six-figure budget to start. You need consistency.
Will this reduce AI hallucinations?
Ideally, yes. If platforms face liability for errors, they will improve their systems. But hallucinations are inherent to how large language models work. They will never disappear completely. The goal is to minimize their frequency and impact on your brand.
Check where your brand stands in AI discovery. Get your free AI Visibility Score in 60 seconds at audit.searchless.ai.