3 ms·

No one should be liable for anything they don't publish. You can write whatever you want into a private Google doc. Likewise, I can feed whatever prompt I want

No one should be liable for anything they don't publish. You can write whatever you want into a private Google doc. Likewise, I can feed whatever prompt I want into ChatGPT and read the output. By contrast, if I go and publish those false statements, yes I do think I should be liable for that!

notahacker 4y ago
But ChatGPT isn't a private Google document, it's a low temperature content generation model which can emit variations on the same false allegation to a very large number of users if it happens to be relevant to a commonly asked about topic. Clearly this occurring would have the same effect on the reputation of the subject of the false allegation as libel output by a traditional publication. Your suggestion is that culpability for the libel rests with hundreds of individuals that asked innocent questions and received answers, and maybe even shared them whilst reasonably believing them. My suggestion is that OpenAI, the entity behind the model which actually generated the false allegation, might be the more appropriate party to target (with appropriate protections if they respond swiftly to try to prevent the libellous claim from reoccurring)
stubish 4y ago
And this is an example of why the people creating ChatGPT will end up holding the bag. When you read a document produced by ChatGPT, no matter if you prompted it or someone else did, ChatGPT is providing you information. That information can be complete bollocks. If those lies portray a person or group in a bad way, that is defamation in most jurisdictions, because your opinion of that person or group has been negatively affected, falsely. No matter if you choose to republish the information or not. The only fig leaf the creators of these systems has to hide behind is a disclaimer, and they won't hold up in many courts of law.

notahacker 4y ago

But ChatGPT isn't a private Google document, it's a low temperature content generation model which can emit variations on the same false allegation to a very large number of users if it happens to be relevant to a commonly asked about topic. Clearly this occurring would have the same effect on the reputation of the subject of the false allegation as libel output by a traditional publication. Your suggestion is that culpability for the libel rests with hundreds of individuals that asked innocent questions and received answers, and maybe even shared them whilst reasonably believing them. My suggestion is that OpenAI, the entity behind the model which actually generated the false allegation, might be the more appropriate party to target (with appropriate protections if they respond swiftly to try to prevent the libellous claim from reoccurring)

stubish 4y ago

And this is an example of why the people creating ChatGPT will end up holding the bag. When you read a document produced by ChatGPT, no matter if you prompted it or someone else did, ChatGPT is providing you information. That information can be complete bollocks. If those lies portray a person or group in a bad way, that is defamation in most jurisdictions, because your opinion of that person or group has been negatively affected, falsely. No matter if you choose to republish the information or not. The only fig leaf the creators of these systems has to hide behind is a disclaimer, and they won't hold up in many courts of law.