My answer to a recent MSE question was deleted by a moderator a few hours ago. I have posted the screenshots below for those below 10,000 reputation. Yes, part of it did technically use AI to generate content, but it was very clearly for the purpose of showing why AI is useless, and very clearly stated as AI. Are we really not able to recognize that while this may be a violation of the letter of the GenAI policy, this is clearly not a violation of the spirit of it? Does this really necessitate deletion? If so, why?
-
3I imagine the post was getting continuously flagged for violating policy. At some point moderators must bow to the will of the community.– Robert LongsonCommented Jul 15 at 13:57
-
1Then, why was it continiously flagged is my question@RobertLongson– StarshipCommented Jul 15 at 13:58
-
4Because it violates policy as I already explained. Lots of people are fed up with ai generated posts and I we leave yours up people will say why can't my posy stay if that one is OK.– Robert LongsonCommented Jul 15 at 13:59
-
But how does it? Yes it may violate the letter of the policy, but clearly the policy wasn't meant to prevent an answer like this from existing, but rather to prevent someone spamming SE with bad AI generated answers/questions. And a reason of deleting because there is good reason to delete a similar post but not this one is entirely beuracratic nonsense @RobertLongson– StarshipCommented Jul 15 at 14:00
-
6At a minimum, it doesn't follow the requirements in How to reference material written by others, as is mandated in Code of Conduct: Inauthentic usage policy, so is, at least technically, a violation of the Code of Conduct, even though there's intent to indicate it's not your content. I have no insight into why it was actually deleted, but that it wasn't even in compliance with the baseline rules applicable everywhere on the network is a reasonable starting point, but doesn't account for specific rules which exist for content here on MSE.– MakyenCommented Jul 15 at 14:03
-
1Of the 3 things I am suppossed to do, I should have provided a link (but I wish someone could have just said that in comments or something) and I did the other 2.@Makyen– StarshipCommented Jul 15 at 14:07
-
It was also formatted to look like a totally AI generated reply, though don't know if intentional or not. I justify the deletion, even if I wouldn't have voted to delete it myself.– Shadow WizardCommented Jul 15 at 14:10
-
1@ShadowWizard it wasn't totally AI generated, and it definitely wasn't meant to look like it. I thing I was pretty clear on what was and was not AI generated in the answer.– StarshipCommented Jul 15 at 14:11
-
10Your post already had a perfectly fine comment explaining why it should be deleted, which you aren't showing in your screenshot here.– Tinkeringbell ModCommented Jul 15 at 14:29
-
No, it doesn't. It links to said policy, which I explained about. And I said that I would remove the AI content anyways...@Tinkeringbell– StarshipCommented Jul 15 at 14:32
-
6Ah, so there was an explanation and a comment.– Journeyman GeekCommented Jul 15 at 14:33
-
Yes, there was an explanation that didn't make sense (why should violating a letter of this policy be cause for deletion when there are lots of other options). @JourneymanGeek– StarshipCommented Jul 15 at 14:40
Add a comment
|
3 Answers
I can't explain why the answer was deleted, so obviously all of us non-mods are acting, to some degree, in speculation here, but... I don't think the AI-generated blurb was necessary. You could've had a substantial answer to the question without it at all. I recognize that you were using it to demonstrate a point, but you yourself admit that it does, technically, violate the policy, so I'm a little confused why you're confused that it was deleted. You just pasted the AI response without commentary to, effectively, pad your answer. That wasn't the intent, I understand that, but that's the end result. My recommendation is to just not post AI-generated content.
A quick remedy to this would be to edit out the AI-generated content and flag the answer for a moderator to undelete it. You could put the AI-generated content into a blockquote and provide commentary about why portions of the output are bad, but I really don't think the juice is worth the squeeze here. The point you're making can be much more convincingly explained using your own words, and citing sources that explore the accuracy/reliability of AI tools in the context of technical questions.
-
I would be willing to remove the AI content (as that is clearly what people want here), but I was wondering why it was deleted. I'm confused why it was deleted by our mods are people capable of recognizing what a rule was meant to do, and what is a false positive– StarshipCommented Jul 15 at 14:13
-
1There's generally no false positives. There's exceptions where the genAI output is essential to illustrating a point and used in moderation. Commented Jul 15 at 14:16
-
"There's generally no false positives" - By a false positive, I mean something violating the letter of the policy that I don't think (and I doubt you do either) the policy intended to prevent. @JourneymanGeek– StarshipCommented Jul 15 at 14:20
-
8@Starship one of the main things we wanted to avoid with the policy is having AI crap on the site. And your answer was bringing AI crap to the site. Yes, you said it was AI, and sure, you did it to prove a point, but the end result was a net increase in the amount of AI crap on the site. Even in the screenshot here, that was too long to bother even reading. I don't see any benefit in including it and I can see the downsides.– terdonCommented Jul 15 at 14:32
-
Okay. I still don't understand why downvotes/comments cannot be used for the function of saying that you think the AI content isn't being helpful. Clearly at least 2 people did think that answer was helpful. Or the content could have been edited out. The remainder of the answer was helpful (methinks). @terdon– StarshipCommented Jul 15 at 14:34
-
5I voted to undelete, now that you removed the AI bit, @Starship, but please also fix the formatting. You are using quote blocks as headers, which is confusing (I though you were quoting the OP at first) and hard to read. And note the point the deleting mod made in their comment: you hadn't even quoted the actual quote, the bit that came from the AI.– terdonCommented Jul 15 at 14:38
-
-
5Big? Perhaps not, but it is semantically wrong (so will break screen readers) and it is confusing to the reader since you are using a thing for something that isn't what the thing is designed to be used for.– terdonCommented Jul 15 at 14:42
I'd initially posted an answer that contrasted a GenAI produced result to mine (produced by experimentation and human stubbornness), but It felt off. I self deleted, rewrote the answer to be entirely mine, and left my prompt in place as an exercise to the reader. It's difficult to integrate genAI output in a valid way into an answer in most cases, and I decided the effort wasn't worth the potential benefit.
With respect to meta - and especially main meta - in this particular case the rules as written and intended seem valid.
I think a critical thing here is - it wasn't acting as critique or an example, it was just padding. You posted genAI output, and I don't see any particular specific use of it to illuminate the person asking a question or educate. Even there, its a blurred line which honestly, I had trouble keeping to my satisfaction on my answer.
I often talk of effective meta as storytelling. What was the story you were trying to tell with that chunk of genAI? It was a very ineffective, one might even say weaker half of your post.
answered Jul 15 at 14:03
-
1The story I was trying to tell was look at AI answers, they aren't really that useful. The point is that much of it was off-topic and some stuff entirely fictious, and provided little to no good answer. @Journeyman Geek– StarshipCommented Jul 15 at 14:06
-
1And while it was, your answer didn't effectively illustrate that Commented Jul 15 at 14:09
-
Yet, again, why did it need to be deleted though? Not everything that didn't illustrate something well needs to be deleted. Downvoting, editing, and commenting exist.@Journeyman Geek– StarshipCommented Jul 15 at 14:11
-
6Well there's a rule saying that you shouldn't post gen AI content did and you did. While I believe in flexibility, that would need an exceptional reason. If I was a moderator, and hadn't posted an answer (conflict of interest) I'd have considered the same Commented Jul 15 at 14:14
-
"that would need an exceptional reason" - Is this is clearly not the use of AI the policy is meant to target a good reason? Again, why delete rather than edit it out or something like that.– StarshipCommented Jul 15 at 14:16
-
7See, you just posted a chunk of AI with minimal context. It was deleted because its a large chunk of your answer, and moderators can't read your mind. At some point, its your responsibility to fix your answer. You clearly know its due to the use of AI output, and its against the rules here, so.. its up to you to fix it. Commented Jul 15 at 14:19
See here:
What counts as “content generated by generative artificial intelligence tools”?
Answer: (emphasis mine)
“Content generated by generative artificial intelligence tools” is any content crafted, in part or in whole, using a tool that writes a response automatically based on a prompt it is provided.
So, strictly speaking, even small part of a post that was generated with AI can justify deletion.
answered Jul 15 at 14:20
-
1No one is disputing that it violates the letter of that policy. But we aren't idiots and we can exercise the spirit of that policy.– StarshipCommented Jul 15 at 14:22
-
4@Starship so if you agree it violates the policy, what exactly are you asking? I really don't understand. Commented Jul 15 at 14:27
-
1Why we can't think about the spirit of said policy and prioritize that over the precise wording. @Shadow Wizard– StarshipCommented Jul 15 at 14:28
-
3@Starship because then we'll open a wide gap. Too wide. When something is against the policy, delete. Simple and effective. If the author can later edit and make it valid, they can flag, it can be undeleted. Still simple, and no harm done when all sides understand. Commented Jul 15 at 14:35
-
1Is it not opening a wide gap that we can't ever ignore the letter of the polcy in favor of it's spririt?– StarshipCommented Jul 15 at 14:41
-
3@Starship no. You can say, though, that it makes the place to be very strict, which some might consider as "not welcoming", and new users seeing it can be scared enough and not even post - i.e. it drive away potential users. That is valid point and might be true, but overall, a fair exchange. Better have strict policy that keeps high quality. All that said: it's really up to the site mods. As I already mentioned in other comment before, I wouldn't delete it myself - but I do accept the mods deciding to delete it, that's all. Commented Jul 15 at 14:47
-
1My point is that there are times (maybe you don't think is one of them but they exist) where the letter of a policy prevents a clearly good answers/questions which it wasn't designed for, in which case it should be ignored.– StarshipCommented Jul 15 at 15:15