7 ms·

It has nothing to do with monopolies. Google was protected from defamation law with search because the page title and snippets were direct quotes from the linke

It has nothing to do with monopolies. Google was protected from defamation law with search because the page title and snippets were direct quotes from the linked result page. Whereas with AI overviews, the copy is written by a Google-controlled LLM.

andai 4mo ago
Making the response consist entirely of direct quotes sounds like a better user experience than what they're doing now.
Terr_ 4mo ago
I recall "Large Libel Models" was one of the sobriquets going around. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4546063 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4546063
- smcin 4mo ago
  Brilliant
Hfuffzehn 4mo ago
Yes, the monopoly is not relevant for the court. It is relevant for Google though, because they want to transfer it to another product. And the court is saying that whatever that new product is, Google is not allowed to mislead the public by pretending it is search.
- fauigerzigerk 4mo ago
  I'm not sure. Is it really just the misleading part that the court takes issue with, i.e portraying Gemini output as a search result? In my view, the ruling could mean that Gemini's output is legally seen as first-person speech by Google regardless of where it is published.
  - Hfuffzehn 4mo ago
    Sure that could be the worst case. But then we are in the case law vs. continental law debate for lawyers. When I wrote "Gemini is not illegal" it would have been more correct to say that the court has not decided about Gemini but specifically about search. And I do think it makes sense to distinguish between explicitly chatting with an LLM and entering something in the search bar people have used for decades now.
    - fauigerzigerk 4mo ago
      I completely agree with you that this particular ruling does not make Gemini illegal in Germany. I'm just wondering whether the logic of the ruling is tied to the specific context (search results) or applied more generally to LLM generated text.
  - Ekaros 4mo ago
    I think that will depend. If person publishes it as their own speech Google is likely not fully liable. On other hand as Google publishes it as their own speech on their own site well they are fully liable. To me it seems pretty reasonable that content Google generates with Gemini and publishes on their own google.com domain is their speech. Or at least they are fully editorially responsible.
  - danaris 4mo ago
    Based on the article, the problem isn't that Google is attempting to present the AI Overview as "search" to the users. Google is attempting to claim to the court that AI Overviews are just like search, and the court isn't buying it. The problem here is that 1) the AI Overview is giving incorrect information, and 2) it is Google's own words ...which makes Google liable for anything false there. They weren't liable for false statements in search results, because that was 100% other people's content, being linked to, with a small blurb taken directly from their words. With the AI Overviews, regardless of whether it's getting its information from other people's content, it's remixing it based on Gemini's algorithms and synthesising statements that are frequently wrong.
    - 4mo ago
Saline9515 4mo ago
It should be noted that defamation law has a very low threshold in Germany, to the point that businesses routinely sue Google Maps users for less-than-5-stars reviews. Google had to change their display of reviews because of this. Three stars review is taken down for "libel": https://support.google.com/maps/thread/367778263/google-maps-ratings-and-reviews-becoming-increasingly-useless?hl=de https://support.google.com/maps/thread/367778263/google-maps... HNer gets a legal threat after saying that he didn't like a doctor: https://news.ycombinator.com/item?id=44734895 https://news.ycombinator.com/item?id=44734895 Google maps german policy: https://support.google.com/contributionpolicy/answer/16997273?hl=en https://support.google.com/contributionpolicy/answer/1699727... Meanwhile, service in Germany is still rather poor (especially if you have children), but at least no one can complain!
- dgellow 4mo ago
  Damn, that makes public facing LLMs a horrible liability
  - Blackthorn 4mo ago
    Good.
  - pif 4mo ago
    Which is what thay should be in any sane environments. LLMs can be wonderful tools, but they cannot be trusted.
  - tempfile 4mo ago
    Public-facing LLMs are a horrible liability.
  - InitialLastName 4mo ago
    No more of a horrible liability than a human writing public-facing copy.
- applfanboysbgon 4mo ago
  > HNer gets a legal threat after saying that he didn't like a doctor: https://news.ycombinator.com/item?id=44734895 https://news.ycombinator.com/item?id=44734895 This post is written by an LLM, and possibly a complete fabrication.
conartist6 4mo ago
Right. If Google isn't liable for that content nobody is.
- tavavex 4mo ago
  That's probably what they actually think at Google. A lot of the "they're just like people" marketing now seems like responsibility-washing. That despite Google developing the LLM and charging for its use, they treat it less like a piece of software and more like providing access to some guy's answers who has no agency and isn't owned by them. If their software ever messes up and calls you a murderer to a potential employer, or offers dangerous advice, their reaction would be "oh well, that's just too bad isn't it? Well, we wrote that the LLM may not always be right in page 683 of the service agreement, so this is not our problem."
panarky 4mo ago
So if the actual page is defamatory then Google quoting it verbatim is protected. But if Google accurately summarizes the defamatory page, then the summary is defamatory?
- deanishe 4mo ago
  Did you read the article? The whole point is that these particular defamatory claims weren't made by some third-party, but hallucinated entirely by Google's own AI. There is no "accurate summary". Just a pack of lies fabricated in-house at Google.
- brainwad 4mo ago
  It could presumably accurately summarise it while still sticking to reported speech. In German there is a grammatical tense that is basically only used for this; it's pops up in newspaper reporting a lot where they don't want to endorse the statements they are reporting.
nonethewiser 4mo ago
So AI isnt just ripping off the underlying information?

andai 4mo ago

Making the response consist entirely of direct quotes sounds like a better user experience than what they're doing now.

Terr_ 4mo ago

I recall "Large Libel Models" was one of the sobriquets going around. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4546063 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4546063

smcin 4mo ago
Brilliant

Hfuffzehn 4mo ago

Yes, the monopoly is not relevant for the court. It is relevant for Google though, because they want to transfer it to another product. And the court is saying that whatever that new product is, Google is not allowed to mislead the public by pretending it is search.

fauigerzigerk 4mo ago
I'm not sure. Is it really just the misleading part that the court takes issue with, i.e portraying Gemini output as a search result? In my view, the ruling could mean that Gemini's output is legally seen as first-person speech by Google regardless of where it is published.
- Hfuffzehn 4mo ago
  Sure that could be the worst case. But then we are in the case law vs. continental law debate for lawyers. When I wrote "Gemini is not illegal" it would have been more correct to say that the court has not decided about Gemini but specifically about search. And I do think it makes sense to distinguish between explicitly chatting with an LLM and entering something in the search bar people have used for decades now.
  - fauigerzigerk 4mo ago
    I completely agree with you that this particular ruling does not make Gemini illegal in Germany. I'm just wondering whether the logic of the ruling is tied to the specific context (search results) or applied more generally to LLM generated text.
- Ekaros 4mo ago
  I think that will depend. If person publishes it as their own speech Google is likely not fully liable. On other hand as Google publishes it as their own speech on their own site well they are fully liable. To me it seems pretty reasonable that content Google generates with Gemini and publishes on their own google.com domain is their speech. Or at least they are fully editorially responsible.
- danaris 4mo ago
  Based on the article, the problem isn't that Google is attempting to present the AI Overview as "search" to the users. Google is attempting to claim to the court that AI Overviews are just like search, and the court isn't buying it. The problem here is that 1) the AI Overview is giving incorrect information, and 2) it is Google's own words ...which makes Google liable for anything false there. They weren't liable for false statements in search results, because that was 100% other people's content, being linked to, with a small blurb taken directly from their words. With the AI Overviews, regardless of whether it's getting its information from other people's content, it's remixing it based on Gemini's algorithms and synthesising statements that are frequently wrong.
  - 4mo ago

Saline9515 4mo ago

It should be noted that defamation law has a very low threshold in Germany, to the point that businesses routinely sue Google Maps users for less-than-5-stars reviews. Google had to change their display of reviews because of this. Three stars review is taken down for "libel": https://support.google.com/maps/thread/367778263/google-maps-ratings-and-reviews-becoming-increasingly-useless?hl=de https://support.google.com/maps/thread/367778263/google-maps... HNer gets a legal threat after saying that he didn't like a doctor: https://news.ycombinator.com/item?id=44734895 https://news.ycombinator.com/item?id=44734895 Google maps german policy: https://support.google.com/contributionpolicy/answer/16997273?hl=en https://support.google.com/contributionpolicy/answer/1699727... Meanwhile, service in Germany is still rather poor (especially if you have children), but at least no one can complain!

dgellow 4mo ago
Damn, that makes public facing LLMs a horrible liability
- Blackthorn 4mo ago
  Good.
- pif 4mo ago
  Which is what thay should be in any sane environments. LLMs can be wonderful tools, but they cannot be trusted.
- tempfile 4mo ago
  Public-facing LLMs are a horrible liability.
- InitialLastName 4mo ago
  No more of a horrible liability than a human writing public-facing copy.
applfanboysbgon 4mo ago
> HNer gets a legal threat after saying that he didn't like a doctor: https://news.ycombinator.com/item?id=44734895 https://news.ycombinator.com/item?id=44734895 This post is written by an LLM, and possibly a complete fabrication.

conartist6 4mo ago

Right. If Google isn't liable for that content nobody is.

tavavex 4mo ago
That's probably what they actually think at Google. A lot of the "they're just like people" marketing now seems like responsibility-washing. That despite Google developing the LLM and charging for its use, they treat it less like a piece of software and more like providing access to some guy's answers who has no agency and isn't owned by them. If their software ever messes up and calls you a murderer to a potential employer, or offers dangerous advice, their reaction would be "oh well, that's just too bad isn't it? Well, we wrote that the LLM may not always be right in page 683 of the service agreement, so this is not our problem."

panarky 4mo ago

So if the actual page is defamatory then Google quoting it verbatim is protected. But if Google accurately summarizes the defamatory page, then the summary is defamatory?

deanishe 4mo ago
Did you read the article? The whole point is that these particular defamatory claims weren't made by some third-party, but hallucinated entirely by Google's own AI. There is no "accurate summary". Just a pack of lies fabricated in-house at Google.
brainwad 4mo ago
It could presumably accurately summarise it while still sticking to reported speech. In German there is a grammatical tense that is basically only used for this; it's pops up in newspaper reporting a lot where they don't want to endorse the statements they are reporting.

nonethewiser 4mo ago

So AI isnt just ripping off the underlying information?