There’s been a lot of talk about AI getting things wrong. They call these hallucinations, where an AI system presents information that is wrong or made up. It’s only recently that the legal question of who bears responsibility for those false statements was answered.
This June, a German court made a landmark ruling that could reshape how we think about AI, search and corporate accountability online.
A regional court in Munich found Google directly liable for harmful false statements generated by its AI Overview function. The court explained precisely why Google can’t hide behind the traditional liability shield that has protected search engines in the past.
For anyone working in digital marketing or SEO, it’s important you’re aware of this outcome. It signals a shift in how the law views AI-generated content, and what that means for how information appears in search results.
The case
It was centered around German publisher, Verlagshaus24, which found itself being falsely linked to fraudulent schemes and subscription traps through Google’s AI Overview function.
When users searched for the company name alongside terms like “fraud scheme,” Google’s AI was generating summaries that connected the publisher to suspicious business practices, despite no such connection existing.
The publisher sent Google a cease-and-desist letter, but Google didn’t respond appropriately. The false statements remained and it was taken to court.
The significance of this case isn’t just that Google lost, it was the court’s reasonings for ruling against Google and what it reveals about how AI-generated search results should be treated.
Search engine defence doesn’t apply anymore
For years, search engines have enjoyed a degree of legal protection. Mainly because the links and snippets they display from websites across the internet aren’t adopted as their own, they are just pointing people towards it. In their eyes, they are not responsible for the accuracy of third-party content.
Courts accepted this logic realising that the burden of verifying every single result would be near to impossible. Search engines wouldn’t be able to operate as we know it without some form of protection.
But the German court drew a critical distinction. Google’s AI Overview isn’t a traditional search result in the sense that it doesn’t just display links or brief snippets. It independently compiles information from multiple sources, weighs that information according to its own algorithms, and then generates a summary statement in its own words. Essentially, it creates new information as opposed to finding information that already exists.
It’s for this reason that the court said Google couldn’t claim the same liability protections as an ordinary search engine. The AI Overview is Google’s own statement, made by Google’s own AI, offered to users as an answer. And these overviews show for almost 55% of Google searches. So, when an answer contains false information, Google is wholly responsible.
Why this matters for businesses
Now it’s confirmed that false information appearing in AI Overviews is a legal liability for Google, they have a concrete incentive to improve the accuracy of their AI summaries, otherwise they face legal consequences.
But there is a broader implication that should matter to anyone concerned with their online reputation. The court’s reasoning suggests that any AI-generated content that appears under a company’s name and authority is the company’s responsibility. You can’t blame the AI and wash your hands of liability.
This is a timely reminder that what appears in your brand’s search results carries legal weight, whether it’s in the form of traditional results, featured snippets or AI Overviews. If it’s not addressed, the appearance of false information could damage your brand and bring legal liability for the company displaying it.
AI hallucinations
Although the ruling was conclusive, it doesn’t solve the problem of AI hallucinations. It doesn’t prevent Google’s AI from making things up or drawing connections that don’t exist. But it does create accountability when it happens.
AI systems, particularly large language models, have a tendency to assert things that aren’t true. They do this because they’re pattern-matching systems trying to provide helpful, coherent responses – not because they’re checking facts against a reliable source. The technical community is aware and has been working on the problem, but it’s far from solved.
Information is extracted from various sources to provide an overall conclusion, which often ends up far beyond what any single source said. In this case, it created a false connection between a legitimate publisher and fraudulent schemes. That connection didn’t exist in any of the source material and was created by the AI’s processing.
The concerning factor
If you’re running a business and your competitors, or unrelated disingenuous companies, are mentioned alongside you in Google’s AI summaries, you now have legal recourse. Whilst it’s good, it’s also reactive. It is down to you to find the false statement and take action, which may require legal proceedings to correct it.
The real challenge sits with prevention. How do you stop your business from being falsely associated with fraud, scams or unethical practices in the first place? This remains unclear, but we do know that the quality of information about your business online holds legal importance. This includes what’s written about you, where it appears, and how AI systems interpret it.
What comes next
Whilst the German ruling isn’t UK law, European courts tend to influence each other. It’s likely we’ll see similar cases and rulings elsewhere.
It still raises questions about how AI systems should be regulated, who’s responsible for AI-generated content, and whether search engines (or other companies deploying AI) need stricter controls on what their systems can claim.
If your business is being falsely described in AI-generated search summaries, you have grounds to challenge it. And if you’re in the business of deploying AI systems that make claims about people or companies, you’re liable for the accuracy of those claims.
It’s a significant shift, yet only the start of how the law will grapple with AI accountability.
