-6

My initial understanding is that Stack Exchange currently bans ChatGPT generated answers.

But is it allowed to include text that ChatGPT has generated as part of a question?

As an example:

Suppose I have some question about history, and I try to answer it using ChatGPT. ChatGPT gives a plausible answer, but gives sources that refer to some book without specifying the page.

Upon closer inspection, it seems that the information isn't in the book.

Then, is it allowed to ask a question on SE for a reference for a historical event, possible verification or refutation of the generated answer, and include what output ChatGPT gave, with the attribution that it was generated by ChatGPT?

Improve this question
7
  • 8
    Different sites have different ChatGPT rules so I suspect this isn't really generally answerable for all sites. ChatGPT is well known to create plausible garbage though so asking to explain that garbage is just a waste of time.
    – Robert Longson
    Commented Feb 20, 2023 at 17:26
  • 17
    As always, this is just an elaborate XY problem. You asked ChatGPT for X, gave you some output Y, you now ask on SE for the Y. Instead of straight up for the X.
    – VLAZ
    Commented Feb 20, 2023 at 17:36
  • 7
    I'm surprised that it gave you the title of a real book instead of a fake one (or a real title with a fake author). Either way, the chance of the book actually containing what Chat GPT said it contains is basically zero and you should stop relying on it for factual information in any form.
    – Laurel
    Commented Feb 20, 2023 at 17:49
  • 8
    It's already been said that each site has their own rules surrounding ChatGPT. Some sites also frown upon pupils cheating by posting homework here. "Find me a source because ChatGPT doesn't" sounds a lot like someone trying to do their homework using ChatGPT, and thus may be frowned upon as academic dishonesty too.
    – Tinkeringbell Mod
    Commented Feb 20, 2023 at 18:18
  • For any claim you see anywhere, if you do not know whether it is true you can research & post an on-topic non-duplicate question re its truth. Do not post more than one question in a post. Dumping more than one claim to be fact-checked is asking more than one question & lack of research effort merits downvotes. ChatGPT is suitable only for drafting reasonable phrasing & generating claims for you to recognize as true & use, to recognize as false & not use, or to not know the truth of & to research. It is not appropriate to post a dump of text from anywhere & request fact-checking en masse.
    – philipxy
    Commented Feb 21, 2023 at 2:14
  • physics.meta.stackexchange.com/q/14281/123208 has examples of bogus references generated by ChatGPT.
    – PM 2Ring
    Commented Feb 21, 2023 at 23:10
  • This is an example of an Android end-user question from ChatGPT. As VLAZ mentioned, it becomes an XY problem. Instead of really asking "Is there ChatGPT app on Android?", it actually becomes "Can ChatGPT be set as Personal Assistant on Android?"
    – Meta Andrew T.
    Commented Feb 22, 2023 at 16:34

2 Answers 2

Reset to default
7

This seems like documenting your research:

Search, and research

Have you thoroughly searched for an answer before asking your question? Sharing your research helps everyone. Tell us what you found and why it didn’t meet your needs. This demonstrates that you’ve taken the time to try to help yourself, it saves us from reiterating obvious answers, and above all, it helps you get a more specific and relevant answer!

so I do not see any reason for it not to be allowed as long as it's properly indicated as such (which you intend). For sites with a blanket ban, you could just paraphrase the content. I see people talking about the results of their Google/Wikipedia/... searches, or asking a friend/colleague/teacher all the time - there's no reason to treat AIs differently. On average, it may produce worse results than a search engine (depending on the topic and your search skills) right now - the situation will no doubt be different in 6-8 months/quarters/years. I mean, if even Google Alphabet is scared about it ...

Now, whether this counts as sufficient research, that's another question which probably varies per site. Some sites have explicit close reasons for this reason, on other sites users might just downvote you.

We recently had a similar discussion on English Language Learners: Is it on topic to ask about English advice given by ChatGPT?. For another example, see Has checkmate ever been achieved with the move 'castles kingside' in a high level chess game? on Chess.SE.

Stack Exchange currently bans ChatGPT generated answers.

FWIW, that's not true; some sites do, but many sites allow them provided they are properly attributed and when users do not blindly copy-paste the output but review it thoroughly. Unfortunately, 99% of the users don't ...

Improve this answer
3

ChatGPT is not banned on Stack Exchange (SE), ref. Ban ChatGPT network-wide, ChatGPT is banned in some SE sites. What is not allowed in SE is plagiarism. Ref. How to reference material written by others.

To learn in which sites there might be a ban you might start by looking at Is there a list of ChatGPT discussions and policies for our sites?, but if you already have a clear idea of what you want to ask and know on which site that is on-topic, the best is to go directly to the corresponding per-meta site and look for the most recent discussions about the policies related to the use of AI generated text content.

Regarding the example provided, there are a lot of sites with questions with the tag [scope] (this tag is used in many per-site meta) that included the word "history" : [scope] history is:question returns 284 questions.

Related

Improve this answer

Not the answer you're looking for? Browse other questions tagged .