17

The SE network has done some testing on a "No AI content allowed" banner and has made it available for sites that request it.

We've had quite a number of AI generated answers here lately and I think this might slightly reduce the number that we see in the future. It certainly won't make it worse.

Each site has to request it, so I'm proposing that we do so.

I guess that an Up Vote on the question indicates agreement that we should move that way? (I dunno - there really isn't a poll option here...)

13
  • I don't understand the whole thing. If it's correct then why does it matter. It's the humans clicking on the arrows that are the problem. But seeing as there's no solution for that, then yes(?), assuming AI answers are total crap. Would we ignore an AI script that auto closed dupes? Kinda hard to drive ad traffic if this is just all computers doing SE the way it's supposed to be done. 10 years-in there are no new questions. - That happened at like year 3. We just keep turning over users who know that and/or care.
    – Mazura
    Commented Jan 7 at 20:00
  • 3
    Will the banner help? Probably only a little, because the people actually posting the AI junk probably don't care much. But hopefully the banner will discourage some otherwise well-meaning "hey, I found a ChatGPT answer to this question!" people. Now if an answer deleted because AI despite the banner could trigger an automatic long-term ban of the user then we might be on to something...
    – manassehkatz-Moving 2 Codidact
    Commented Jan 7 at 22:53
  • 5
    @Mazura - this has been debated to beyond blue-in-the-face on the more 'computer-oriented' stacks & main meta. Amongst the troubles with AI 1) It doesn't check its answers & makes things up with equal authority as accepted fact [google 'AI hallucinations']. 2) it is plagiarised from uncredited sources.
    – Tetsujin
    Commented Jan 8 at 18:19
  • 3
    The use of AI, @Mazura, eliminates the need to think. Either on the part of the person posting it (copy question, paste into AI random generator, copy result, paste into answer), or on the part of the reader (the computer says it's true, so it must be true). People have a hard enough time thinking for themselves as it is, I'm against encouraging less thinking.
    – FreeMan
    Commented Jan 8 at 18:24
  • 5
    Experiments were done in the 80s. Kids were given calculators and allowed to use them for math tests. Some gave correct answers, some gave answers that were off by 10%, some off by 50% (numbers approximate - I don't remember the exact details). Kids just wrote down the answer the calculator gave them, even when it was blatantly wrong, like 5 x 10 = 55. The math questions were appropriate for the age group and were simple enough that the kids should have been able to do them in their head and known that the calculator was wrong. It's not like they were asking 6 year olds to do calculus.
    – FreeMan
    Commented Jan 8 at 18:27
  • 3
    I suggest some extra points - "No AI questions, No AI comments, no AI voting" Because making it clear seems wise
    – Criggie
    Commented Jan 15 at 18:15
  • 2
    I'm with you 100%, @Criggie. Can we get SE to modify the message for us? :D
    – FreeMan
    Commented Jan 15 at 18:16
  • 3
    @Criggie - I think you may have to raise a separate request This one as it stands has been flagged to staff as our intent to proceed, by being tagged 'status-review' by a mod. I presume when done, it will be re-tagged by staff as 'status-complete'. Right now there are only two minor variations on offer & I never considered the 2nd option to be appropriate, so I very much assume we'll get the main one. Bear in mind, this will appear when someone tries to answer, so it won't necessarily be the appropriate place to mention comments or votes.
    – Tetsujin
    Commented Jan 15 at 18:49
  • Having no idea whatsoever how the status-review process works or what kind of time frame there might be, could someone please explain and/or give an update on where this is in that process? Will we be notified if this is approved or denied?
    – FreeMan
    Commented Jan 18 at 16:27
  • 2
    Update from SE: meta.stackexchange.com/q/396626/163027
    – BMitch Mod
    Commented Feb 9 at 15:56
  • 2
    I've also listed this post here: meta.stackexchange.com/q/384922/163027
    – BMitch Mod
    Commented Feb 11 at 14:24
  • Great updates, @BMitch. Is there any expected timeline for getting the banner posted here? I would presume that everyone will see it at least once, so we'll know once it's in place, I'm just curious on whenish.
    – FreeMan
    Commented Feb 12 at 12:42
  • 1
    @FreeMan they have a meta post describing the process and their historical response rate, but not a lot of detail on the timeline. Running the SEDE query with the last 3 months in the range, I see a lot of sites asking for the banner and none are marked as completed yet.
    – BMitch Mod
    Commented Feb 12 at 16:51

3 Answers 3

Reset to default
3

The banner for no AI content allowed has been turned on. Apologies for the long turnaround time.

2
  • Better late than never! Odd, though, since SE is going to their own AI, but hey, what the heck...
    – FreeMan
    Commented Apr 16 at 15:22
  • 1
    @FreeMan Have no fear. They can eventually change it from "Don't use AI" to "Don't use free AI. But sign up and use our premium AI for a small monthly fee."
    – manassehkatz-Moving 2 Codidact
    Commented Apr 17 at 4:35
14

I'm just going to copy/paste my answer from Super User's Meta on the same topic…

Yes.

It can't hurt. It costs us nothing. People will either read it & pay attention or not, but if only 1% read it & then don't post the AI answer they were about to, then the banner has been successful, even if just to a small degree.

Even for those well-versed in site etiquette, you'll only see it once, so the annoyance factor is low.

1
  • Agree. Human answers are on-brand for StackExchange. Announcing that with zero ambiguity for answer consumers is also useful.
    – popham
    Commented Jan 14 at 21:28
0

I don't think we really need this alternative answer, but let's see.
As noted by FreeMan in the OP, Meta doesn't have a true poll structure other than for Moderator Elections, so the accepted method is to provide two opposing answers to a binary decision & count those up/downvotes. (& as mentioned in comments, downvotes don't do my rep any good, so I'm hoping people will just upvote their preferred answer & ignore the other ;)

This separates votes on agreeing/liking or disagreeing/disliking the question itself to just votes supporting each argument [as meta is not the same as main, a downvoted question may or may not indicate a bad question but a disagreement of the sentiment it proposes].

No.

We don't need this.
[Supply your own reasoning if necessary.]

4
  • 1
    You may want to make this a community answer so the down votes don't hit your personal score. OTOH, you may not care (I mean, how many rainbows and unicorns does a guy need) but I thought I mention it.
    – FreeMan
    Commented Jan 8 at 18:33
  • 1
    I hadn't considered that, tbh, but if I make one CW, then I ought to make both CW. Now I'm well past 10k on here, I'm no longer worried about the odd downvote here & there. I could hope people will only upvote their preferred answer rather than downvote the other. We'll see ;))
    – Tetsujin
    Commented Jan 8 at 18:35
  • 3
    I was gonna down vote this one (just to offset any crazy people :D), but you seem nice enough, I didn't want to give you a bad rep. ;) But, fair point, I won't vote on this one at all...
    – FreeMan
    Commented Jan 8 at 18:38
  • 4
    @FreeMan Meta doesn't have actual rep so no votes here affect your main site rep
    – Machavity
    Commented Jan 9 at 19:09

You must log in to answer this question.

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