Google’s AI search told users that an Aachen doctor had been convicted of a sex crime. The conviction belonged to a different man, and on September 24 the Aachen Regional Court [in Germany] banned Google from repeating it, ruling that the company answers for what its AI Mode says. It is the second German court this year to treat Google’s AI answers as Google’s own words, and the first ruling to say so has now been settled rather than tested on appeal.
…
The doctor asked Google to delete the text on August 28 and followed up with a formal legal warning on September 15. Neither worked. The court’s 8th civil chamber then acted without hearing Google first, citing the urgency of false claims that were “publicly visible” to an unlimited audience. Google Ireland, which runs Google Search in the EU, now faces a fine of up to €250,000 for each breach, or detention of its managers, if it repeats three versions of the text. Google can still challenge the order.
The court’s reasoning is blunt: Google “provides AI-generated search and is therefore also responsible for the answers generated in this context.” It added that checking would have been easy: had Google “confronted its own AI” with the doctor’s objections, it would have found the statements were false.
…
That reasoning comes almost word for word from Munich. On May 28 the Munich Regional Court I ruled (case 26 O 869/26) on AI Overviews that told searchers two Munich publishing companies were “known for dubious business practices”, ran subscription traps and kept changing their names to avoid being traced. None of the linked sources said that; the first one cited did not mention the companies at all.
Google argued that it was only summarising other people’s pages, the same defence that has long protected search engines in Germany. The court disagreed. AI Overviews, it said, do not just show links but make “independent, new and substantive statements” built by evaluating and combining third-party sites, so Google must treat them as its own, including as the provider of an AI system under the EU’s AI Act. “Only it has influence over the AI,” the judges wrote.
Google said at the time it would appeal. That appeal has ended in a settlement, The Decoder reported on Saturday; the terms are not known and neither Google nor the publishers would comment. The upshot is that no higher court has ruled, so the Munich judgment is not binding precedent, but lower courts such as Aachen are already following it.
…
ChatGPT can’t discuss Brian Hood because he sued them (it shuts down if you bring him up).
We should all have the option to opt out of being discussed by LLM models. In fact, it should probably be an opt-in process. Not being discussed should be the default.
The sooner we make companies responsible for their products the better.
Honestly there should be a lot more Brian Thompsons by this point.
Fuck anyone who speaks ill of Luigi
For a company earning $403 billion in annual net income, €250,000 is just another budgeted operating expense. Fines should always be based on annual worldwide revenue.
Detention of its managers might have more impact.
In before managers become minimum wage contract gig economy workers. Gotta Weasle word proof it somehow.
That’s exactly what needs to be done
Well good news then I guess.
Eh, when you consider how often LLM’s spit out wrong answers it could add up quick.
Is that supposed to be a downside? Seems like the consequences of not producing accurate and fully tested products that have the potential for massive negative effects
No, it is not a “downside”, it is why this court ruling matters.
Fines should actually impact companies, not just be another cost of doing business. By imposing a high dollar value fine per violation, regardless of frequency of violation, companies may actually be incentivized to stop breaking the law in order to continue existing
There’s a difference for a one-time fine that covers continued wrong doing over a period of time, and a fine about a specific instance of misbehaviour.
The first don’t accrue, and I agree, should be based on revenue so that they serve as actual punishment and serve as future deterrent.
But a €250k that can be slapped by thousands of cases that happen every day (AI providing wrong information) has the potential to be massive as well.
I think the point other people are making is that this can set precedent in German law and allow other people with similar issues to bring suit and win on merit and not have to fight as hard for their case to be considered. It could also end up inspiring similar cases in other countries using similar frameworks, adapted for local laws.
If suddenly Alphabet has thousands people taking to them to court in jurisdictions all over the world, that can indeed hurt them, although it will take longer than injunctions on Google now, I fully agree.
Now if only everybody had the same funds as that doctor and could simply sue Google.
Make it a recurring subscription instead.
10% of yearly revenue global for making a machine that lies (in a legally-compromising way) once per year. You want a lie per month? That will be 120% of your revenue for that year.
The numbers can be tweaked, but it should not be tolerated. Were I to misbehave and say what Google’s “AI” said, I would be in much more trouble than that.
Stupid fucking comment. OOOO The LLMS wrong answers could add up quick, better let them sling slop and ruin people’s lives. Fucking dipshit.
?
I fail to see how you got the idea I support LLMs or their companies from my post.
Uncapped indexed fines is the tastiest thing any country could implement. I wonder why whole world is not running on them.
Why is anyone listening to AI? We were told not to use Wikipedia for reports…wtf
We were told not to use Wikipedia for reports
Yeah maybe in 2003.
Yeah, pretty much that exact window. Before that it was dont trust ANYTHING on the internet.
Don’t trust anything on the internet should have really never gone away.
I fully agree, but the thing here is this doctor who was accused of a sex crime he didn’t commit. They can’t simply throw out a chatbot without a human being to be held accountable. This is one of the major threats of LLMs in my opinion: the danger isn’t the LLM itself, but its creator’s attempt to reject any human responsibility.
Deny justice long enough and you shouldn’t be surprised when the people extract their own form in the worst way possible.
Its the same as the SaaS revolution. The accountability is outsourced and everyone can just shrug. Same as police violence. Homeless camp gets disrupted and now there is a flower patch…huh. If no one is responsible, than no one has to be responsible.
Maybe I am immune to responsibility and I just haven’t tried it out yet.
“Let me google that” no longer is actually searching for an actual article sadly (even if only the 1st), but just parroting whatever its LLM vomits out.
Because wikipedia was never backed by billionaires owning virtually everything
If Google was updating Wikipedia and falsely accusing somebody of sex crimes on their Wikipedia pages, I think you could expect that Google would get sued in that situation, as well. It’s really about the source of the information.
The upshot is that no higher court has ruled, so the Munich judgment is not binding precedent, but lower courts such as Aachen are already following it.
This sentence sure does read like the author is glad the judgement isn’t binding, as opposed to lower courts following it.
“upshot” doesn’t mean, or even imply, that it’s a good thing. It just means something “result” or “consequence”
