-7

Stack Overflow recently released its policy on GPT. There are also some implications in the writing that the rationale is well beyond GPT and software. In general, such answers should be fact-based

https://stackoverflow.com/help/gpt-policy

While GPT makes a lot more sense to be Stack Overflow-centered, is the implied policy also valid on all other Stack Exchange websites?

For example, should pure opinion-based comments and answers without related backing be accepted? Note, my question goes beyond GPT although GPT could be used for any website (I believe).

6
  • 2
    meta.stackexchange.com/questions/384396/…
    – Andy
    Commented Dec 11, 2022 at 6:25
  • 1
    OK, so that covers the GPT aspect. What about the rationale aspect? I assume that is valid for each community Commented Dec 11, 2022 at 6:30
  • 8
    Well - in most cases, those machine generated answers are nonsense and we can deal with them appropriately in the absence of a broader policy. Commented Dec 11, 2022 at 6:34
  • 1
    And are policies like The community trusts that users are submitting answers that reflect what they actually know to be accurate and that they and their peers have the knowledge and skill set to verify and validate those answers. valid between all sites? Commented Dec 11, 2022 at 6:41
  • 13
    Well, it seems a natural thing to post answers you know that work, in a replicable fashion, as opposed to just copy pasting random stuff if the off chance that it works Commented Dec 11, 2022 at 7:35
  • 7
    "pure opinion-based comments and answers without related backing" are already not accepted in the entire network.
    – VLAZ
    Commented Dec 11, 2022 at 8:28

1 Answer 1

0

In the absence of a network wide policy, the same conversation about how to handle AI content is happening in a lot of different places, and as a result the exact reasoning and policy may vary from site. However, I think the majority of sites are concluding that they don't really want AI answers on their sites. (And some sites might not even be having a conversation — there are some sites where AI generated content hasn't been a problem at all.)

I don't think sites that deal with subjective content are going to be more accepting of AI answers. A lot of sites formed their policies around the blog post Good Subjective, Bad Subjective. To summarize, a good subjective answer is backed up, usually by personal experience.

Here's an example policy on subjective Q&A (written years ago for a world with no elegant AI writing), from RPG Meta: How do we ask and answer subjective questions?. Answers on RPG are backed up either by quotes from the rules (objective answers) or by some "when I played this game…" type of experience. For some reason, AI doesn't seem to be able to do the former at all (copyright?) and AI has never played an RPG so it has no way to back up an answer with the latter (except by lying, I suppose). The one AI answer I saw on RPG that I saw was bad enough that I think it would have been deleted even if nobody on the site knew it was AI. (It didn't seem to realize that the question was being asked about a specific RPG.)

As for one of my own sites, the other mods were pretty excited about the Stack Overflow help center article (which I had to remind them wasn't available on our site). That article ended up being one of the references I linked to when writing our policy but otherwise wasn't a big influence when I was writing out my reasoning. You can see the policy here on English Meta: Announcement: AI generated answers are officially banned here. Some of the reasoning has come from the discussions elsewhere (e.g. AI content is tricky to attribute) but some of it is unique to the site (knowledge of English correlates with proficiency in writing, but not when AI is doing the writing).

Like I said, not all sites have had any discussion. The prudent thing to do would be to start the discussion yourself on that site's Meta before posting an AI answer.

You must log in to answer this question.

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