Who actually bears responsibility when the AI responds?
Why the first judgment on AI search results matters far beyond Google
Generative AI is changing the way we search for information.
Instead of clicking through numerous search results, users often now receive a ready-made answer. Search engines condense information, evaluate sources, and formulate their own texts. This saves time.
At the same time, however, a new question arises: Who is liable if these answers are wrong?
The Munich I Regional Court had to deal with exactly this question. At the center was Google’s „AI Overview“ feature, which falsely linked a publishing house to fraudulent schemes and questionable business practices. The court made it clear that this was not a mere reproduction of external content, but an independent content created by Google.
AI not only summarizes – it creates new statements
The crucial difference lies in how generative AI works.
While traditional search engines primarily link to external websites, AI systems formulate their own summaries. In doing so, information from various sources is analyzed, linked together, and presented in a new linguistic format.
This resulted in statements that were not included in any of the original sources. In the Munich proceedings, this led to false accusations being made against the companies involved.
The judgment does not only affect search engines
Even though the lawsuit was directed against Google, its significance extends much further.
More and more companies are using generative AI, for example for
The more independently AI content is formulated or evaluated, the more important the question of who is responsible for its content becomes.
„The AI wrote it“ does not protect against responsibility
In practice, we often encounter the following statement: „The AI wrote that.“
However, this statement does not relieve companies of their legal responsibility.
Anyone who publishes AI-generated content or uses it in front of customers, employees, or business partners should first verify its accuracy.
Especially when it comes to factual claims, erroneous statements can have significant legal consequences, regardless of whether they were formulated by a human being or an AI.
The most common misunderstandings
1. The AI only returns existing information:
Generative AI creates new formulations and can also create irrelevant connections in the process.
2. Responsibility lies with the AI provider:
Anyone who publishes or uses AI-generated content remains regularly solely responsible for their actions.
3. Users already recognize when an AI is wrong:
Especially liquid-formulated answers often seem particularly convincing, even if they are factually incorrect.
What companies should now consider
From our consulting practice, it is particularly advisable to,
Conclusion
The judgment of the LG Munich I concerns Google at first glance.
In fact, however, it raises a fundamentally more important question: Who assumes responsibility when artificial intelligence not only reproduces information but also forms it independently?.
Our experience shows that this very question is becoming increasingly important for companies.
Because the more naturally AI is used for research, summaries, or decisions, the more important it becomes not to adopt its results without verification, but to question them critically.
FAQ
Who is responsible for false AI-generated content?
According to the judgment of the Munich I Regional Court described in the article, an provider of AI-generated content can be held responsible if the system does not merely reproduce external content, but also formulates independent statements. In the specific case, the court deemed Google’s „AI-powered overview“ to be its own content, which Google is responsible for.
Why are AI answers to be evaluated differently from classic search results from a legal perspective?
Generative AI not only links to existing sources, but also analyzes information, connects it together, and forms new texts from it. This allows for statements that were not included in any of the source materials.
Does the ruling only apply to search engines like Google?
No. The article also applies the question to other areas of generative AI use, such as internal knowledge databases, chatbots, customer communication, decision support, or automated summaries.
Can companies claim that „the AI wrote it“?
Companies that publish AI-generated content or use it with customers, employees, or business partners should verify its accuracy beforehand.
What are the typical misunderstandings in handling AI content?
An unnecessary survey is processing without a valid basis. This can lead to regulatory measures and fines, and open the door for claims from those affected. What is most relevant in practice is the effort involved: the more unnecessary fields you create, the greater the obligations for providing information, deleting data, and documenting the data.
What should companies consider when using AI-generated content?
Companies should review AI content before publishing, verify factual claims with reliable sources, establish release responsibilities, and sensitize employees to hallucinations and misinformation. AI should be understood as a support for professional audits, not as a substitute for them.
