The interesting thing about the German court ruling against Google is not the verdict.
-
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.
@david_chisnall Google is just one of the many big tech companies that I have become disillusioned with. Gemini will literally lie to you and then gaslight you about lying to you. Creating an AI with the sole purpose of being "likable" based on RLHF was never a good idea. It will LITERALLY lie and make crap up just to "sound informed," without actually providing any factual information.
When you do ask it to teach you something beneficial, such as topics that are STEM related like ethical hacking, it will tell you "this is unethical and I can't help you with this" knowing damn well that real threat actors aren't relying on their app to craft precision engineered payloads or that script kiddies aren't going to spend an hour debugging a payload that is only going to work in a vuln-by-design sandbox. It has "this is just to generate revenue" written all over it.
"Just in: a script kiddie sends an unobfuscated payload to port 1337 for several fortune 500 companies, goes completely undetected, and now the authorities can't track them down." - said nobody ever
There's so much good that could come from such powerful tools and they are literally choosing to run us into the ground with them. It's covert warfare against us IMO.
-
@david_chisnall It's like when Microsoft said Copilot was for entertainment purposes only and everybody immediately stopped using it for serious applications... wait, no, they didn't.
AI boosters are just gonna rationalize it as "well they HAD TO say that, it was just a court strategy, they don't really MEAN it" and nothing's gonna change.
@ratsnakegames @david_chisnall
The point being made wasn't that AI boosters will or will not be convinced, the point being made is that there is now legally un-retractable evidence that Google is liable for fraud if they advertise LLMs as accurate and someone buys into the advertisement and Bad Things Happen. The "someone" can, after the fact, sue them for misrepresenting the capabilities etc.
-
@DJGummikuh @david_chisnall "Google's lawyers" is doing a lot of heavy lifting here! As is always the case, there will be mutiple different law firms working for Google in different legal systems and jurisdictions, with different understandings of their customer's priorities and needs. Obviously somebody drafting that submission didn't realize that Google is an AI firm first, not an advertising firm or a search firm, and they were undermining their customers' highest priority.
@cstross @DJGummikuh @david_chisnall I think it's more likely Google know their court filings are never going to hurt their public image, because the number of people whose minds might possibly be changed by them is tiny. There is not much daylight between "See! Google admit their AI is useless!" and "Google will say whatever they need to to win a case, just like every other corporation out there."
-
From the decoder's writeup:
At the hearing, Google argued that users could check the linked sources themselves to verify whether the AI summary was correct. Users generally knew "that information generated with AI should not be blindly trusted," the company claimed. That's a remarkable statement given the scale at which Google serves AI overviews. It's also not entirely true, since the connection between sources and generated content isn't always there.
The court rejected this. The possibility of disproving a statement through further research doesn't "regularly exempt from liability for this statement." The AI overview was "understandable on its own" and contained "a self-contained statement with independently understandable content and no reference to other possible interpretations or even unreliable content." Studies show that users almost never click on sources in AI overviews, which supports the court's reasoning.
-
@syn @david_chisnall Tucker Carlson is off of Fox now, although still doing the same shit somewhere. I would say that the distinction is that while there are loads of right-wing ideologue grifter types in his mold and millions of dollars for them, no one’s trying to prop up the entire economy on a shell game based around right-wing grifter personalities.
@c0dec0dec0de
so you're saying one Tucker Carlson is much easier to remove than one "AI overview" ?
@syn @david_chisnall -
@robinsyl Yet. We'll see how well that holds up in court the first time someone loses money or reputation as a result of something Copilot does.
Fully employment for lawyers coming up!
@david_chisnall @robinsyl remember when llms were supposed to put lawyers out of their jobs
-
@ratsnakegames @david_chisnall
The point being made wasn't that AI boosters will or will not be convinced, the point being made is that there is now legally un-retractable evidence that Google is liable for fraud if they advertise LLMs as accurate and someone buys into the advertisement and Bad Things Happen. The "someone" can, after the fact, sue them for misrepresenting the capabilities etc.
@eschwartz that is neither an accurate summary of the original toot nor of the legal situation
-
@cstross @DJGummikuh @david_chisnall I think it's more likely Google know their court filings are never going to hurt their public image, because the number of people whose minds might possibly be changed by them is tiny. There is not much daylight between "See! Google admit their AI is useless!" and "Google will say whatever they need to to win a case, just like every other corporation out there."
@freakazoid @cstross @david_chisnall the point, as pointed out in the OP, is that this now admissible evidence in further trials. Fuck public opinion, this admittance of Knowledge is going to be exploited to the extreme on the european attempts to reel in the overreaching power US AI companies have into european markets
-
@freakazoid @cstross @david_chisnall the point, as pointed out in the OP, is that this now admissible evidence in further trials. Fuck public opinion, this admittance of Knowledge is going to be exploited to the extreme on the european attempts to reel in the overreaching power US AI companies have into european markets
@DJGummikuh @cstross @david_chisnall Fingers crossed!
-
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.
@david_chisnall
Google's famous slogan, "Don't be evil," should probably be updated to: "Don't be as evil as we are."
#Google #LLM #Liability #Ai -
@c0dec0dec0de
so you're saying one Tucker Carlson is much easier to remove than one "AI overview" ?
@syn @david_chisnall@wolf480pl @syn @david_chisnall I think I’m agreeing that the economic implications are more likely to have an effect when they’re applied to a technology that cannot be trusted rather than a person who cannot be trusted (even if that person is emblematic of a large cohort of similar grifters). While Fox News claim that its evening shows are entertainment and contain no actual information could be made to stick to the individual hosts, Google admitting that LLMs aren’t fit for purpose could meaningfully be attached to the whole category of technology and not just Gemini (that’s Google’s Lying Language Mangler, right?).
-
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.
@david_chisnall wait.... so google actually rocked up to the court with the fox news 'no reasonably intelligent person would mistake out programming for news, and if you do you are a fucking moron' business model??
-
From the decoder's writeup:
At the hearing, Google argued that users could check the linked sources themselves to verify whether the AI summary was correct. Users generally knew "that information generated with AI should not be blindly trusted," the company claimed. That's a remarkable statement given the scale at which Google serves AI overviews. It's also not entirely true, since the connection between sources and generated content isn't always there.
The court rejected this. The possibility of disproving a statement through further research doesn't "regularly exempt from liability for this statement." The AI overview was "understandable on its own" and contained "a self-contained statement with independently understandable content and no reference to other possible interpretations or even unreliable content." Studies show that users almost never click on sources in AI overviews, which supports the court's reasoning.
@david_chisnall It's a little bit too complicated for now, so never mind.
-
@eschwartz that is neither an accurate summary of the original toot nor of the legal situation
@eschwartz Yes, Google going on record and admitting they know LLMs are fallible MAY make them liable to false advertising charges in the future, as David points out. But they can just add some weaselly legalese to their ads, à la "Copilot is for entertainment purposes only", and people will still believe the bullshit, despite the disclaimer saying outright that the bullshit is bullshit.
The impact will be very limited, is my point.
-
@david_chisnall that's absolutely fascinating, the argument is somehow that it is a work of complete fiction which is derived from non-fictional sources(assumed). I guess we are back to the impossibility of validation. Does this stop them cold? Or do we have to wait until other countries test the same case.
@david_chisnall further, I would wonder whether they could simply change their format to provide references with excerpts from those references. I can hope.
-
@eschwartz Yes, Google going on record and admitting they know LLMs are fallible MAY make them liable to false advertising charges in the future, as David points out. But they can just add some weaselly legalese to their ads, à la "Copilot is for entertainment purposes only", and people will still believe the bullshit, despite the disclaimer saying outright that the bullshit is bullshit.
The impact will be very limited, is my point.
@eschwartz And yes, there is a line where you can no longer waive liability with a disclaimer. I'm sure Google will manage to find exactly where that line is.
-
@cstross @DJGummikuh @david_chisnall
Meanwhile in the guitar world, Fender Guitars’ lawyers and upper management are commiting brand suicide.Is there a reason these cases are being tried in German courts?
@JamesPadraicR @cstross @DJGummikuh @david_chisnall because US courts are hopelessly corrupt? Those suits wouldn't stand a gnat's chance in a hurricane here in the USA, unfortunately.
-
@cstross @DJGummikuh @david_chisnall
Meanwhile in the guitar world, Fender Guitars’ lawyers and upper management are commiting brand suicide.Is there a reason these cases are being tried in German courts?
@JamesPadraicR @cstross @DJGummikuh @david_chisnall Maybe because Germany has something that looks like rule-of-law?
-
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.
@david_chisnall Meh, they'll just announce that their next model is so much more advanced that the claim no longer applies. "Oh no, you see, that was for LAST month's LLMs..."
-
@david_chisnall It's like when Microsoft said Copilot was for entertainment purposes only and everybody immediately stopped using it for serious applications... wait, no, they didn't.
AI boosters are just gonna rationalize it as "well they HAD TO say that, it was just a court strategy, they don't really MEAN it" and nothing's gonna change.
@ratsnakegames @david_chisnall like when any psychic medium performs a stage show they need to say (in the UK at least) that they’re not claiming to be able to talk to the dead and that their performance is for entertainment purposes only. Yet more often than not the audience members fully believe the entertaining performer is a genuine spiritualist with a connection to the other side and wouldn’t have bought tickets otherwise. The boilerplate is passed off as a legal necessity, at odds with the overall impression being created and sold.
