Inspired by this post:
It should be against the rules to post AI crap without putting it in a quote-box. There’s enough AI crap on the internet that this board should, IMO, make this minimum effort to make it easy to identify (and thus ignore, if so desired).
Continuing the discussion from What were you THINKING?:
This is an excellent idea and should get some discussion outside the BBQ Pit. To repeat for emphasis.
Proposed rule:
If you’re quoting something not written by yourself, put it in a quote box. This could be text from another poster or copied from another source, such as another website or an AI chatbot.
I merged the two threads together.
Moderating:
Sounds good in theory but how would mods catch it? AI is becoming more and more convincing and less obvious than before. It can sound really human-conversational, to the point some humans may get accused of using AI when not quoting AI, or vice versa, AI slips in unnoticed.
Yes, I agree. I just created a thread for the same point (I’ve flagged it to be merged. And so it has.)
Most of my irritation of AI content would be soothed if it were in a quote box. We do that for quoting other websites. We should do it for any significant block of text coming from a source not the poster.
Co-sign.
I previously advocated that AI slop should be inside a “hide details” collapsing section thing, so anyone who wants to see it can open it but it’s not inflicted on the rest of us, but got no traction. If I can’t have that, I’ll grudgingly accept a quote box as a compromise.
Mods won’t always catch it, just like they won’t always catch socks and trolls and other rule violators. That doesn’t mean those are bad rules.
I don’t think what msmith537 wrote was inappropriate: they did say the response was from ChatGPT and they did put it inside quotation marks (though they may be hard to notice.) That said, it would be helpful to use properly attributed quote blocks for AI responses. (Ideally, as someone who generally supports AI, I would think it is more useful for a poster to read the response and form their own reply based on it, as AI slop does annoy me on message boards. If I want a ChatGPT response, I’ll use ChatGPT. I want to hear what the personalities of the Dope think in their own words, even if some of it is informed by AI research.)
Very short snippets could be in quote marks, but anything longer than about a sentence or two should be in a quote block. Visually, a paragraph should be in a quote block.
Agreeing; with the addition that the post should include saying who the quote’s from, with link if applicable.
ETA: of course this won’t stop deliberate rule breakers, any more than any other rule does. But I think a lot of the problem isn’t from people wanting to sneak AI in undetected, but with people who just want to use it without clearly distinguishing it from their own post.
I would very much like this to be the case, still a distant second to “don’t copy/paste AI because this is a board for humans to read things written by other humans.”
I don’t know that it should be in a quote box, per se. But it should be clear that you are posting something from AI somehow, whether it’s a quote box, or italicized, or in a collapsed spoiler box.
I also don’t think we really need a new rule, do we? Isn’t it already forbidden to copy-and-paste from elsewhere without attribution? If I were to find a social media post, or a news article, or a YouTube video, and reproduce what was there as if it was my own words, wouldn’t that already be a violation? It should be, if not.
The only thing I’d propose as a new rule for AI contributions is to require people to say where it’s from. And not “AI”, but rather which AI. Because they aren’t all the same, and different AI engines have different reputations. Much like you shouldn’t say “I read this in the news”, it matters if this was from the New York Times or National Enquirer or One American News Network or Russia Today.
Can someone link to the offending post that set this off, if any rules are being added or appended, this has to be presented to @Ed_Zotti.
I just merged the 2 almost simul threads on this topic and don’t know the root cause. I know it didn’t happen in the What were you THINKING? - #6737 by Alessan thread.
Warning, Pit Language, so I’ll dropdown and blur, because people may not want to read the initial exchange - Again Pit Language and insults:
Me, I’d actually prefer a spoiler box so I know going in not to waste any of my time, but a quote box is acceptable.
That is someone being angry at another poster, but it points to here:
Is that correct?
We wouldn’t need a new rule, no, we have an existing one, which this would be a modification of.
Artificial intelligence (AI). AI-generated responses must be clearly marked as such. Do not use AI to generate entire posts. AI responses are often unreliable – if you use them as the primary source for your assertions, you’re likely to get a lot of flak on this board.
@what_exit, I accidently linked to the insult, hit the back link to see the post that started it. I’ll go correct.
ETA - Initial AI post now corrected, FTR, I think said poster did a reasonable job of confirming the source, though they apparently didn’t know the rule cited above and it did end up making nearly the entirety of the post.
So it sounds like the rule already exists. Maybe it could use the suggestion that it either be in a Quote Box saying which AI or detail spoilers saying AI text.
I’m not sure @msmith537 's post is or isn’t clearly marked. He wrote:
“I asked ChatGPT “is collecting things the sign of a disturbed mind” and here is what it responded” so probably counts as clearly marked.
On the other hand it is a long, probably too long quote from ChatGPT. I don’t care much about this issue and even I think that would have been far better hidden by the Detail spoiler tag.
If he was quoting anything not public domain, he would get a modnote about quoting more than fair use allows. I have no idea about ChatGPT rules and fair use. @Ed_Zotti made it clear in the rules that “if you use them as the primary source for your assertions, you’re likely to get a lot of flak on this board.”
So rule of thumb, is use AI minimally. Make sure you present it as such.
I agree. Also seeing some Facebook posts (also possibly AI) reproduced with no indication they’re not the poster’s words. A general reitteration that anything you didn’t write needs to be attributed and clearly marked would help.
I have no problems with at least making it a “best practices” option, especially in more serious threads. The rule as stands says “clearly marked” and this would be a good way to do it, though of course, the final call is in @Ed_Zotti’s hands.