Topic: Shouldn't there be a forum rule against obvious AI-generated content? (Read 1202 times)

I don't think moderators are leaving cases where it can be proven that someone used an AI bot to write a post unmoderated. The information from this thread already shows that the staff is deleting such posts. The consequences we aren't seeing are bans for users who post AI-generated content. At least not on a wide scale.

Very little if you ask me. If I felt like talking to bots, I would do that and not be on Bitcointalk. I would much rather speak to real people and understand what they think of different ideas and concepts.

I reported him to Hhampuz yesterday, and he removed him from his campaign with immediate effect as soon as he saw my PM.



He also posted an apology in the thread where he was reported for plagiarism which I showed above but he did it by saying that he didn't want to use AI to generate and make posts but he was using it for translation and AI added extra stuff itself in the translations, lol.

What kind of person would get something done and then don't check if it is done in the correct way or not? He probably thinks we are all dumb or something by making such excuses for what he did. 

This cheater is so stupid and frankly deserves to be punished accordingly. Can't stand such people. I'm pretty sure no signature campaign manager would tolerate such blatant plagiarism and abuse of LLM content, especially not @Hhampuz.

It can't be categorized as spam because spam is different. A spam is a post or text that doesn't have any meaning, value, or constructiveness in it whereas AI-generated content tends to have meaning and usually is on-topic as well if the user knows how to use it properly. It falls more under the category of plagiarism than spam.


It's understandable, that the decreasing activity metrics might be a bit concerning, but is that enough reason to leave certain cases unmoderated? Wouldn't that make things worse? I wonder what value a user brings to this forum who knows nothing more than copying content from elsewhere and posting it in the forum as their own whether it's taken from an AI model or somewhere from the internet.

I don't know but in my opinion, less activity but a cleaner forum is much better than more activity but a forum full of spammers, plagiarizers, and AI shitposters.

---

This is what I mean when I say "obvious" AI-generated content:


The original post is still not deleted but he has removed the part highlighted in the report above.

What's surprising is that the user in question, Sg4j1n3ll0, is a participant of Duelbits Signature Campaign managed by Hhampuz:



So basically, he is not only doing a very bad thing by posting AI-generated content and passing it on as his own, but he is also cheating a signature campaign and getting paid for it.

Should this be tolerated or ignored only because it might reduce forum activity if he gets punished? I don't think so.

Moderators and admins might have become more lenient towards banning plagiarized content altogether. We don't know. They are not going to tell us that they will be tolerating some cases of plagiarism. Why do I think that?

Two reasons:
1. The forum isn't as active as it was in the past. Issuing more bans will only decrease the activity further. 
2. I see no ban appeal threads in Meta dating back to mid-August. Usually, there are at least a few. I only checked using the keyword "appeal."

The forum isn’t one that changes easily and quickly. Before any chnage is been made on the forum although, the sort you are asking isn’t exactly a change but, an update on specifics with regards to rules and developments on the forum.
However, though AI generated content is hardly regarded as plagiarized content, it’s agreeable that they lack originality and don’t relay the opinion of the user. Hence, this counts as spam and could be treated in the same way.

Spams do get deletes accompanied with temp bans in serious cases and after some series of these temp bans, the sword is allowed to swing.

AI-generated content should be reported immediately without a delay, this to me is an advanced way of plagiarism, which is a bit difficult to detect. Practically, this is you claiming to be better than what that you are not really are. Well, in other hands, the victims aren't doing themselves that good as overtime you may become dump, can't become a better version of yourself, it means you can't think outside the box ,and offer something from your inner-self

Exactly. Users who are not banned for posting AI-generated content might take advantage of the fact that moderators are lenient with it and that there is no official rule against it which shouldn't be the case, in my opinion. If plagiarism is punishable, proven AI-content plagiarizers should be punished as well.


It's true, that you can't provide much evidence for AI-generated content because you can't match the post with the content just like how we do with plagiarism because the content is only accessible by the user who generated it, however, if there is clear cut cases where it is very easy to understand that a user is posting AI-generated content again and again, they should get some treatment for it.


That would do, I guess. Even if there is no separate rule for it, having it mentioned in the official rules should be enough for it to be imposed and acted upon when needed.

That's right, as long as AI-generated posts that get reported are deleted by moderators we don't actually have to care about it being a unofficial rule. I'm pretty sure that admins aren't adding it as a rule because in future things will change a lot and there's chance that AI will continue to get improve overtime.

The tools that detect AI content may not be able to detect the content when AI gets more smart and mimics human writing perfectly. We all know that AI content is just plagiarized content and there's actually an unofficial rule against plagiarized content and that's enough to report such posts and so far whenever I've reported any such posts they were deleted by admins/moderators within few minutes to hours and that's a good thing.

I don't find it OK to only delete AI-content that has been presented as own. For me this is plagarism and as such it is supposed to be not tolerated in this forum. The rules for plagarism abuse are there and pretty much clear.

Providing hard evidence for copy/pasted AI-content without proper labelling might be more complicated than simple copy/pastes from "normal" online content. But I would expect same outcome once properly "found and proven guilty" as you say.

Thinking about it again, this forum probably doesn't need a specific rule targeting copy/pasting AI-content, though it wouldn't hurt. The plagarism rule could have an additional sentence clarifying that AI-content without proper labelling it as such is considered to be plagarism, too.

@dkbit98
You are right. I hadn't noticed that the admins actually added a new rule that prohibits the promotion and linking to mixing sites. I went to check after seeing your reply, and it's there towards the bottom.

That shows that they are still making officially unofficial changes to the forum rules when needed. There is no mention of AI-content still. To me, it fits perfectly under rule #33 that prohibits copying content from other sources and websites. AI-generated content is something you copy from an external source/website to paste here and pretend it's your own.

Is it okay if a person is found guilty of plagiarism but as a punishment, only his plagiarized posts are removed and he is left to keep doing that? I have said this repeatedly in this thread, I understand that AI-generated posts are removed, and that is great, but if there is no official rule about it and if it falls under the rule about plagiarism, and if users who get either temporary or permanent ban for plagiarism, the same thing should apply for AI content posters if they are found or proven guilty of it.


It's not only about punctuation marks or correct grammar, there is a pattern most AI models use to generate texts. I use tools such as Grammarly as well to find mistakes in what I write so that I don't have mistakes, but that doesn't make my posts look like they have been generated using an AI model because there will always be a difference in content that you write yourself and rectify using a tool and content that you generate using a tool.

Until they start charging for all AI generating services, and believe me they will do it soon

I agree, and rules can easily be changed, look how quickly they added latest mixer ban rule.
Like I said, use of AI generated tools is fine for me, but only if use of this tools is correctly mentioned.

Plagiarism rule can be applied for AI-generated posts and without any new rule, it's fine with current existing rule for plagiarism.

The content is more important than mistakes in commas, grammatical errors. AI-generated content can be very generic and is not relevant to a discussion but AI usually used by these people to create topics, than normal posts inside a thread.

Do careful checks by ourselves or can use support from Grammarly or AI or similar tools, it's good. Having mistakes in a post is not too good and if we can sharpen our posts with better grammar, more relevant words, it's better.

This use is not plagiarism and corrected contents by Grammarly, AI will be different than contents completely created by AI.

As long as the admins delete AI-generated posts, I don't really care if we have a new unofficial rule added to the list of other unofficially official rules. It's plagiarism, and we already have a rule against plagiarised content.

Someone said you can recognize AI content by the lack of mistakes, commas, and/or grammatical errors. I think it's dangerous to use that as an argument.

I try to pay close attention to how I write my responses and I am sure there are others just like me. People shouldn't have to answer to anyone why there are no or very little mistakes in their forum posts. It's because we try to avoid making mistakes, that's why.

Lazy rich kids always had this possibility: they could pay someone to write their thesis. Now lazy poor kids can do the same.

I know what you mean, so don't worry.
I am just giving an example that content written by AI is not always bad when the information contained in it is correct and valid. But unfortunately, some users don't want to admit that they use AI to post. So again, don't worry because I get what you have been trying to explain.

Bro, we are talking about people using AI-generated content as their own. I don't want to repeat myself again and again, so please read the previous posts in this thread for a better understanding of the context and my post you quoted.

Not always agree. AI content that can be confirmed based on data and facts is also better if you know how to use it. Use quotes and tell everyone that what you are quoting is information gathered by the AI ​​instead of hiding everything as if it were yours. Typo errors made by native English speakers are understandable, but misleading information is also unacceptable.


Here is one example:

When and where did I say using AI or other tools to rectify your mistakes is an offense or should be treated as one? I use Grammarly myself, you can check my post history. Do you think I sound like I'm using an AI to write my posts?

There is a difference between AI-generated posts and posts that have been corrected using AI or tools such as Grammarly. And, this agenda isn't against people who are good in writing English, it's about people using AI and the patterns I have discussed in the opening post refer to that and not to native English speakers.


Just for your reference, I earlier defended a guy for writing original posts but having mistakes in his posts. This proves that I'm never against people who make grammar or spelling mistakes or even words they use as long as I can understand what they are trying to say. I believe that a person who writes original content with mistakes is much better than someone who posts perfect content but generates it using AI or copies it from others.