6 ms·

I don't see how this can be true at all in the search engine context or even chatGPT where you are asking for information and getting back a result which may or

I don't see how this can be true at all in the search engine context or even chatGPT where you are asking for information and getting back a result which may or may not be true. It would be different if an AI is independently creating and publishing an article that has false information.. but that's not the case. You are asking a question and it's giving it's best answer. I'm not a lawyer by any means, so someone please give a more legal distinction here. But if you asked me what the operating margin of company X was, and I give you an answer (whether I make it up or compute it incorrectly), you or the company can't sue me (and win) for libel or anything of the sort. So I'm not sure the potential is as big as you think it is.. that's like saying before any AI you can sue google because they return you a search result which has a wrong answer, or someone just making shit up. That's not on them- it's literally indexing data and doing the best it's algorithm can do. It would only be on the AI if you are literally selling the use of the AI in some context where you are basically assuring it's results are 100% accurate, and people are depending on it as such (there is probably some legal term for this, no idea what it is).

crazygringo 4y ago
> But if you asked me what the operating margin of company X was, and I give you an answer (whether I make it up or compute it incorrectly), you or the company can't sue me (and win) for libel or anything of the sort. If you're a popular website and you intentionally publish an article where you state an incorrect answer that many people follow and make investment decisions about, the company absolutely can sue you and win. In the courts, it will ultimately come down to to what extent Microsoft is knowingly disseminating misinformation in a context that users expect to be factually accurate, regardless of supposed disclaimers. If Microsoft is leading users to believe that Bing Chat is accurate and chat misinformation winds up actually affecting markets through disinformation, there's gigantic legal liability for this. Plus the potential for libel is enormous regarding statements made about public figures and celebrities.
- eatsyourtacos 4y ago
  >you intentionally publish an article where you state an incorrect answer that many people follow and make investment decisions about, the company absolutely can sue you and win. I literally said that in my post! But then I said if you asked me, it's different. You are ASKING the AI to give you it's best answer. That is a million times different than literally publishing an article that people should assume to be factual. >If Microsoft is leading users to believe that Bing Chat is accurate But they aren't, and never will be. So you are basically just making things up in your head for argumentative purposes. There are going to be disclaimers up the wazoo to easily protect them. Partly because, as I keep trying to tell you, it's much different when you ASK a question and they give you an answer rather than publishing something to the public where it's implied that it's been independently fact checked etc.
  - crazygringo 4y ago
    Right, but the distinction of "asking" isn't a legal one I'm aware of. I don't think it matters. If 100,000 people "ask" the same question on Bing and get the same inaccurate result, what's the difference between that and publishing a fact that gets seen by 100,000 people? There isn't one. And Microsoft needs to tread a very fine line between "use our useful tool!" and "our tool is false!" Which I'm not sure will be possible legally, and is probably why Google has been holding back. Bing is clearly intended for information retrieval, not for generating fictional results "for entertainment purposes only", and disclaimers aren't as legally watertight as you seem to think they are.
    - eatsyourtacos 4y ago
      >I don't think it matters. If 100,000 people "ask" the same question on Bing and get the same inaccurate result, what's the difference between that and publishing a fact that gets seen by 100,000 people? There isn't one. Of course there is a difference. Publishing an article is literal intent. The premise is you researched or have knowledge on a topic, you write it, you fact check it, and it's put out there for people to see. An AI which consumes a bunch of data and then tries to be able to respond to an infinite number of questions has no intention of harm doing and you can't even call it gross negligence. It's not being negligent- it's doing exactly what it's supposed to do.. it might just be wrong. I'm not sure in what universe you think those are the same thing. Now if I ask the AI to write a paper about the forecast of a company, and I just take the result and put it into a newspaper where it's assumed it's been fact checked and all that, sure that's completely different. >disclaimers aren't as legally watertight as you seem to think they are I guess you know more than Microsoft's lawyers. I'm sure they didn't think about this at all when releasing it....
adamckay 4y ago
> But if you asked me what the operating margin of company X was, and I give you an answer (whether I make it up or compute it incorrectly), you or the company can't sue me (and win) for libel or anything of the sort. If your answer damages company X then they can sue you. If you Tweet that a vacuum cleaner is terrible because its noisy to your 4 followers it's probably not a big deal as (under UK law, and I'm assuming similar internationally) a company has to prove "serious financial loss". If you write about it on your Instagram that has millions of followers then that's more of an issue, so you can assume a search engine claiming to summarise results but apparently hallucinating and making things up is liable for a defamation suit if it can be demonstrated to harming the company.
- cwkoss 4y ago
  "terrible" and "noisy" are both largely subjective, so it would be very hard to bring a defamation suit in the US over those claims.

crazygringo 4y ago

> But if you asked me what the operating margin of company X was, and I give you an answer (whether I make it up or compute it incorrectly), you or the company can't sue me (and win) for libel or anything of the sort. If you're a popular website and you intentionally publish an article where you state an incorrect answer that many people follow and make investment decisions about, the company absolutely can sue you and win. In the courts, it will ultimately come down to to what extent Microsoft is knowingly disseminating misinformation in a context that users expect to be factually accurate, regardless of supposed disclaimers. If Microsoft is leading users to believe that Bing Chat is accurate and chat misinformation winds up actually affecting markets through disinformation, there's gigantic legal liability for this. Plus the potential for libel is enormous regarding statements made about public figures and celebrities.

eatsyourtacos 4y ago
>you intentionally publish an article where you state an incorrect answer that many people follow and make investment decisions about, the company absolutely can sue you and win. I literally said that in my post! But then I said if you asked me, it's different. You are ASKING the AI to give you it's best answer. That is a million times different than literally publishing an article that people should assume to be factual. >If Microsoft is leading users to believe that Bing Chat is accurate But they aren't, and never will be. So you are basically just making things up in your head for argumentative purposes. There are going to be disclaimers up the wazoo to easily protect them. Partly because, as I keep trying to tell you, it's much different when you ASK a question and they give you an answer rather than publishing something to the public where it's implied that it's been independently fact checked etc.
- crazygringo 4y ago
  Right, but the distinction of "asking" isn't a legal one I'm aware of. I don't think it matters. If 100,000 people "ask" the same question on Bing and get the same inaccurate result, what's the difference between that and publishing a fact that gets seen by 100,000 people? There isn't one. And Microsoft needs to tread a very fine line between "use our useful tool!" and "our tool is false!" Which I'm not sure will be possible legally, and is probably why Google has been holding back. Bing is clearly intended for information retrieval, not for generating fictional results "for entertainment purposes only", and disclaimers aren't as legally watertight as you seem to think they are.
  - eatsyourtacos 4y ago
    >I don't think it matters. If 100,000 people "ask" the same question on Bing and get the same inaccurate result, what's the difference between that and publishing a fact that gets seen by 100,000 people? There isn't one. Of course there is a difference. Publishing an article is literal intent. The premise is you researched or have knowledge on a topic, you write it, you fact check it, and it's put out there for people to see. An AI which consumes a bunch of data and then tries to be able to respond to an infinite number of questions has no intention of harm doing and you can't even call it gross negligence. It's not being negligent- it's doing exactly what it's supposed to do.. it might just be wrong. I'm not sure in what universe you think those are the same thing. Now if I ask the AI to write a paper about the forecast of a company, and I just take the result and put it into a newspaper where it's assumed it's been fact checked and all that, sure that's completely different. >disclaimers aren't as legally watertight as you seem to think they are I guess you know more than Microsoft's lawyers. I'm sure they didn't think about this at all when releasing it....

adamckay 4y ago

> But if you asked me what the operating margin of company X was, and I give you an answer (whether I make it up or compute it incorrectly), you or the company can't sue me (and win) for libel or anything of the sort. If your answer damages company X then they can sue you. If you Tweet that a vacuum cleaner is terrible because its noisy to your 4 followers it's probably not a big deal as (under UK law, and I'm assuming similar internationally) a company has to prove "serious financial loss". If you write about it on your Instagram that has millions of followers then that's more of an issue, so you can assume a search engine claiming to summarise results but apparently hallucinating and making things up is liable for a defamation suit if it can be demonstrated to harming the company.

cwkoss 4y ago
"terrible" and "noisy" are both largely subjective, so it would be very hard to bring a defamation suit in the US over those claims.