This is the script of my national network radio report yesterday regarding a new German court decision holding Google responsible for their AI Overviews misinformation, and the potential global impact of that decision. As always there may have been minor wording variations from this script as I presented this report live on air.

- - -

Yeah, it's finally happened. A court has directly and strongly pushed back against Google's ridiculous excuses about misinformation that spews from their AI. And while this is a ruling from a German court it is so incredibly well reasoned and written that it is very likely to have significant impact on thinking about these issues around the world, and so could ultimately influence other courts everywhere.

We've talked numerous times about AI hallucinations and misinformation, and this particular case involved Google's notorious AI Overviews. Recent studies have suggested that at Google search scale Google is spewing out hundreds of thousands of incorrect AI Overview answers per minute, tens of millions of inaccurate answers per hour. And Google has consistently tried to evade responsibility for these with disclaimers that "AI can make mistakes" and users should double-check the AI answers.

But pretty much everybody knows, and studies have confirmed, that almost nobody tries to verify those answers that Google presents in a very authoritative, "this is the truth" kind of way. I mean seriously, what's the point of getting an AI Overview answer and then having to go digging around researching to try figure out if it's accurate or not? You might as well have done the research yourself in the first place without having your time wasted by an unreliable AI answer that Google itself tells you not to trust!

Also, when Google provides links in their AI Overview answers that connect to the sites that supposedly relate to the answers, it seems that most of the time you can dig through those sites until the cows come home and be unable to find anything that relates to the Google AI Overview answers themselves.

The German court brilliantly notes all these points and various others. The ruling explains that Google isn't just pointing you at sites where you might find useful information related to your search -- as traditional Google search did for decades -- but rather is creating a wholly new authoritative-seeming answer, an answer that cannot be reasonably attributed to anyone but Google itself.

This kind of determination leads us directly into the kinds of arguments that I and others have long been making, that AI answers should not be exempt from the responsibilities that would come into play if those same kinds of authoritative sounding answers had been issued by a human being at the same Big Tech firm in response to user queries.

The kinds of legal protections in place to shield these firms from being liable for third-party content simply do not reasonably apply to first-party created content like AI Overview answers. And this goes far beyond AI Overviews of course, across the entire realm of Large Language Model AI including also AI chatbots, some of which we know have reportedly been implicated in providing advice involved in both murder and suicide cases. New nightmarish instances like these keep occurring, despite Big Tech claims that they're increasing associated safety protocols.

The German court decision holding Google responsible for AI Overview answers is preliminary and there will likely be appeals of course. But the actual reasoning of the decision is enormously solid and persuasive, and may well end up being the first significant crack in the "we don't care" excuse agenda of Google and the other Big Tech AI firms, irrespective of how this specific German court case proceeds.

And from this crack, perhaps we will start to see more moves to actually protect society from AI abuses rather than governments actively encouraging largely if not completely unregulated AI systems. It's going to be a tough slog, because as we know, many politicians in both parties are terrified of the financial and political power of the major Big Tech firms, and even when a politician will privately admit that, for example, they know a massive new data center is going to ruin a community that they represent, they may not have the courage to actually take a stance against it.

So whether or not we'll ever see Big Tech AI CEOs being held personally, criminally responsible for the worst abuses of their AIs -- imagine if you will CEOs being publicly perp walked in shackles to their new abodes in prisons -- at least holding the firms financially responsible for the damages done by their AI systems would be a step in the right direction.

And it appears that the court in Germany, that just effectively told Google to take their excuses for AI misinformation and stuff them, has finally moved the needle in a positive way. It's still a long, long path ahead, and with the untold hundreds of billions of dollars being poured into AI by Big Tech, the battle to protect society will be long and arduous. But perhaps that German court has now exposed a bit of light at the end of that very long tunnel. We shall see.

- - -

L

#google #ai

I finally published my new project, the Banned Book Library! Think of it as a #Cyberpunk dead drop. Reprogram an #esp32 WiFi light bulb to host a web server with books banned in your community. Screw them in wherever is convenient :D. You can't stop the signal!

https://www.richardosgood.com/posts/banned-book-library/

This was partially inspired after reading @benbrown short story "Library" a while back.

https://benbrown.com/library/

• Worüber zu oft gesprochen wird:

0,2% der Bürgergeldbeziehenden sind Totalverweigerer.

• Worüber zu selten gesprochen wird:

0,6% der CDU Bundestagsabgeordneten sind Jens Spahn.

via Laserbrille

Machen wir das zu #streisand

Mehrere hochrangige Wissenschaftler:innen wurden von der Polizei aus einem Fachkongress der American Diabetes Association abgeführt, weil sie das Editorial der Fachzeitschrift Diabetes Care verteilten, in dem die Gesundheits- und Wissenschaftspolitik von Trump kritisiert wird.

Das sind die Leute, die von Free Speech schwafeln.

Die NYTimes berichten (unlocked article):
https://www.nytimes.com/2026/06/05/well/ada-conference-diabetes-trump.html?unlocked_article_code=1.oFA.JZbr.mwpLXkZmrmf9

Das ist das Editorial, das in den USA genügt, um von der Polizei abgeführt zu werden:
https://diabetesjournals.org/care/article/49/6/901/164764/Misguided-Brushes-of-a-Pen-Continue-to-Dismantle

➡️ Update zur Klarstellung: Im verlinkten Artikel steht eh klar und deutlich, dass nicht die Trumpisten selbst die Polizei riefen. Es waren die Organisatoren des Events, nachdem NIH-Direktor Jay Bhattacharya seine an sich vorgesehene Rede zurückgezogen hatte - offenbar wegen der Verteilaktion. Daraufhin gingen die Organisatoren gegen die Autoren des Editorials der Zeitschrift ihrer eigenen Organisation vor.

Das ist mMn noch schlimmer als eine direkte Aktion der Trumpisten. Weil es die Feigheit von Führungspersonen zeigt, die gar nicht zu MAGA gehören, und weil es zeigt, in was für eine Atmosphäre der Angst MAGA die USA geführt hat.

A court in Munich declared that Google is liable for their "AI summaries" and all its hallucinations. This is an important step to bring "AI" slop in line with all other products on the market: "AI" products are basically the only ones where a provider can just deliver unchecked garbage and put all the liability on the consumer. I hope to see aggressive change here.

https://the-decoder.com/landmark-german-ruling-declares-googles-ai-overviews-are-googles-own-words-and-makes-it-liable-for-false-answers/

The interesting thing about the German court ruling against Google is not the verdict. The fact that, if you put libel on your web site, you are liable for it even if you used a machine to automatically generate libel, should not surprise anyone who has paid attention to the law at any point in the last century or so: humans have agency, the tools that they use do not shield them from liability, no matter how obfuscating they are.

The bit I suspect will have much more impact longer term is one of the defences entered by Google's lawyers. Somewhat more verbose in the original German, but it boiled down to: Everyone knows LLMs produce nonsense, no one should ever trust the output of an LLM in any situation that matters, it's not Google's fault if people read the output of an LLM and believed it might have some connection to reality.

It's debatable whether everyone knows that, but this is now an official statement entered into the court record that at least one of the major LLM vendors knows this. And that's now an on-the-record statement made under penalty of perjury that can be entered as evidence in any court case against companies selling LLM-integrated tooling.

I suspect that this will show up in a lot of court cases over the next few years and probably have a much bigger long-term impact than the ruling. Any claim about utility made by vendors of 'AI' tools is now open to lawsuits ranging from misleading advertising to outright fraud as a result of this.

Google would probably have been much better advised to settle the case rather than enter that claim as evidence. Imagine if a car manufacturer had entered a defence against liability in case of a collision by saying 'everyone knows automobiles are impossible to operate safely on the roads and anyone who buys one should know better than to take it on the public highway'. Google's lawyers have just done the equivalent for the 'AI' industry.

EDIT: It hopefully goes without saying, but just in case: I am not a lawyer, this is commentary from someone who watches the industry with a growing sense of disgust, not legal advice.