Germany's Ruling Will Reshape the Internet, Ends "Section 230" Shield for AI
German courts hold Google liable for false information from its AI
Germany has just changed the Internet, ending the “Section 230” Shield companies use to protect themselves from the consequences of their AI.
A court has held Google legally liable for false information generated by its AI search tools, stripping away a critical legal shield the company has relied on for decades. Furthermore, the judge issued a blistering critique of the tech industry's rush to force generative AI into everyday utilities.
The lawsuit stems from a hallucination by one of Google’s AI features when a user was seeking information about a business. It gave the user a response containing false, defamatory or damaging information, and now that business is suing Google. For the past 30 years, Section 230 has protected internet platforms like Google, Facebook, or Reddit from being treated as the “publisher or speaker” of information provided by someone else, thus tossing out all lawsuits.
AI sets that precedent on its head. Because the AI blends, synthesizes, and outputs a unique response, the court decided Section 230 does not apply. Google is now legally viewed as a publisher, making it and all other companies like it directly liable if their AI outputs damaging falsehoods.
Google tried to argue that holding it liable would undermine the utility of AI search and hinder innovation. The judge completely dismissed this defence, effectively stating that "nobody needs AI to search the internet."
The court pointed out that traditional search (using keywords to find a list of index links) worked perfectly fine for decades. The judge argued that tech companies are forcing generative AI onto users not because it inherently improves the core task of finding information, but to chase market trends and maximize ad real estate. Therefore, if Google chooses to deploy an unstable, hallucination-prone technology to replace standard search links, it must bear the legal consequences when that technology inflicts real-world harm.
This changes the entire Internet 🤯!
“AI is experimental” → Disclaimers are legally useless
Meaning:
Companies use disclaimers to waive liability → The court ruled that fine-print warnings do not absolve a multi-billion dollar company from liability for publishing defamation
Google is highly likely to appeal this decision, but for now, the legal armour protecting Big Tech from the mistakes of its own AI has officially cracked. By ruling that AI-generated summaries are original content authored by the platform itself, the court has introduced immense financial risk to companies deploying LLMs (Large Language Models) in public-facing search utilities.
To avoid endless lawsuits, search companies may be forced to scale back AI summaries and go back to simply showing traditional website links.