AI-generated post discussion thread: how to identify & report

[suchmoon]

I've seen ChatGPT spit out some questionable English grammar. You can probably even ask it to make some poorly-Englished texts.

At the end of the day I think we'll lose this battle either way. AI bots will get better, humans will get tired. Ideally mods should have been deleting low value barely-on-topic generic waffle posts anyway. If that were the case then AI wouldn't represent much of an issue. If someone puts in enough effort into coaching the AI to write a good useful text (basically using it to do the google/wiki/etc legwork) that's fair game IMO. But if your regular shitposter just asks AI to write a generic post in this thread then it would stand out like any other sigshitpost and should be deleted.

Keep in mind that we don't need to catch every single shitposter/AI bot/etc. With the evil fee and whatnot it should be enough to just start banning the shittiest ones consistently and make it not worth it for them. I doubt that will happen though, so... then what? Red trust on suspicion of AI usage?

[nutildah]

If someone puts in enough effort into coaching the AI to write a good useful text (basically using it to do the google/wiki/etc legwork) that's fair game IMO. But if your regular shitposter just asks AI to write a generic post in this thread then it would stand out like any other sigshitpost and should be deleted.

That's the thing though: shitposters are turning to AI for the same reason they turn to plagiarism: they are lazy and uncreative. They're not gonna go out of their way to modify the text for those reasons. If they do, then the text stands a higher chance of not being detected as AI-generated, and thus there's no basis for its removal (unless its off-topic / spam).

so... then what? Red trust on suspicion of AI usage?

Nah, in addition to it just bugging me, this thread was motivated by this clause that Royse777 has been putting in his signature campaigns:

What is Prevent AI Domination?
Find members who are using AI text generators, article spinners, plagiarizing forum posting. Report it to me and if the member is in any of my campaign then you will receive the payment for the week instead of the accused member. Depending on the dept of the misuse, the accused member will receive some sort of penalty.

It raises the interesting question of how to create a fair standard for defining AI-generated texts.

Really the only desire I have for shitposters is to be more human and less machine.

[joker_josue]

It raises the interesting question of how to create a fair standard for defining AI-generated texts.

That's what's hard, creating a standard to define what is AI text and what isn't. That's what we shouldn't do, because that could harm users who aren't using those tools. Creating patterns in text is very difficult, because it involves many variables.

Those who misuse the AI will easily be caught, it doesn't take much effort on our part to do so. The biggest problem is, which I don't think exists yet, is the emergence of AI bots that will post on the forum, with their own style, so that the user can level up and participate in campaigns.

[Poker Player]

At the end of the day I think we'll lose this battle either way. AI bots will get better, humans will get tired.

We are already losing it. ChatGPT is being integrated into our lives and work environments in the same way that the Internet began to be integrated 25 years ago and cell phones 15 years ago. We can swim against the tide all we want in this forum to stay in the dinosaur era as the world continues to progress thanks to the internet, mobile and AI.

I am all for AI, especially if it is used as leverage, where the human rewrites and adapts, and not just copy-pastes.

[nutildah]

It raises the interesting question of how to create a fair standard for defining AI-generated texts.

That's what's hard, creating a standard to define what is AI text and what isn't. That's what we shouldn't do, because that could harm users who aren't using those tools.

Just because something is hard doesn't mean it can't be done, or shouldn't be done. Is there even one example of a forum member who was accused of using AI and wasn't? Or that was negatively affected by such an accusation?

Of course its hard to definitively prove one way or the other, which is why I'm not advocating for any witch hunts. But how does one go about providing satisfactory proof that a post was written by AI? I believe its possible to come up with criteria that would satisfy any neutral party and have yet to see any problems caused by it thus far.

Perhaps if a resulting post is considered spam then it doesn't really matter if it was AI generated or not, but for the sake of fulfilling the conditions posed by Royse777 its important to know the difference.

[LoyceV]

At the end of the day I think we'll lose this battle either way. AI bots will get better, humans will get tired.

That is the future I fear: the internet as we know it will cease to exist. For now, the AI-spam is still the result of "ask question, copy result, post result", but in a few years it could very well be: "Hey AI, create a bunch of accounts on Bitcointalk, start posting, join signature campaigns, and make me money!", and by the time that works, it's a lost cause for humans.

[jokers10]

...
I am all for AI, especially if it is used as leverage, where the human rewrites and adapts, and not just copy-pastes.

If you have your own ideas you usually hardly will use AI to express them. Because you can do it by yourself. And if you are too lazy to form your own opinion you'll hardly rewrite or adapt a text. Maybe I'm too pessimistic, but I think that majority of those ones who will use AI for posting on forums will not really take part in discussions or learn something new. AI is good to make work easier but communication with others is IMO not a subject which needs simplification.

[joker_josue]

Just because something is hard doesn't mean it can't be done, or shouldn't be done. Is there even one example of a forum member who was accused of using AI and wasn't? Or that was negatively affected by such an accusation?

I don't want to say, not to investigate and analyze the different cases that arise.

What I am saying is that it will be difficult to create a standard for something that is highly uncertain. Something that has so many probabilities, the creation of a pattern, can affect innocents as well as culprits, and over time it usually tends to affect more innocents.

[Poker Player]

At the end of the day I think we'll lose this battle either way. AI bots will get better, humans will get tired.

That is the future I fear: the internet as we know it will cease to exist.

As if the internet has been with us humans for thousands of years.

The internet practically killed the sending of letters, phone calls from landline to landline, maps, paper newspapers, etc., etc., etc. Related to this, and also to what I commented in the previous post, it has also practically killed the camera, flashlight and diary businesses just to point out a few examples.

When an innovation like this bursts in you have two options: complain and try to stop it (in vain) or try to adapt. I don't know how you see it but the way I see it, if what you say materializes:

... create a bunch of accounts on Bitcointalk, start posting, join signature campaigns, and make me money!", and by the time that works, it's a lost cause for humans.

the difference between a shitposter creating accounts with the AI and you is that you will be able to have them in the top 100 merit earned, while he will not get any. Another thing is that you don't take advantage of the opportunity and you'd rather lament while others take advantage of it.

Maybe I'm too pessimistic,

Yes you are.

[Plaguedeath]

That is the future I fear: the internet as we know it will cease to exist. For now, the AI-spam is still the result of "ask question, copy result, post result", but in a few years it could very well be: "Hey AI, create a bunch of accounts on Bitcointalk, start posting, join signature campaigns, and make me money!", and by the time that works, it's a lost cause for humans.

It already happen since few years ago where people create a bot in order to aim 5000+ followers and many bounty hunters were spamming with their proof of authentication without read the thread, as long as the thread posted in bounties section, it's mean a bounty program even though the subject title and content is talking about token announcement.

It raises the interesting question of how to create a fair standard for defining AI-generated texts.

I don't think there's will be a fair standard because it depends on each user's opinion aka subjective. It will lead a drama as there's no solid evidence, some users will say no and some people say yes.

[karmamiu]

Writing scripts nowadays is very easy with the help of AI, and for that reason even if an AI is used as a tool for writing some well written narratives, there is also an AI that can help you detect narratives constructed by another AI. I have seen several other AI tools that can help detect AI written paragraphs, and most of them are being advertised in tiktok or another social media platform like youtube shorts. I don't know if those AI detectors are really accurate or could really be of help since, I actually didn't and haven't tried them yet and there are too many to mention, some are not even listed on the OPs list (I wish I could list some of them, but I forgot their names), and one of the more reason I am not interested in using those AI detector is that, if AI is the one writing it and detecting, it is no different than the "the ones giving the cure is the one making the disease".

[jokers10]

...
Yes you are.

Okay, can you show anyone who uses AI for posting regularly (not only several times for fun or for research) and is doing it good? All times I saw someone uses it regularly it was a low quality posting with few sense. So if you think that majority will use it in the most right way I guess it is not that I'm too pessimistic, it's about you are too optimistic.

[Obito]

This is pretty tricky, I mean there's probably someone there that doesn't really use AI but gets accused, I think that we need to create a reference to help identify patterns that is consistent with AI usage. Another problem is that what if they use Grammarly or something that corrects their sentences and their punctuations, what happens to them? They might get caught in the crossfire.

[jokers10]

This is pretty tricky, I mean there's probably someone there that doesn't really use AI but gets accused, I think that we need to create a reference to help identify patterns that is consistent with AI usage. Another problem is that what if they use Grammarly or something that corrects their sentences and their punctuations, what happens to them? They might get caught in the crossfire.

I'm sure having "too good grammar" will never be the only reason for blaming anyone in AI posting. It's nonsense to say someone: hey, you are too good so get a punishment. Few sense, non natural text, not answering to others questions, — there could be much more clues to find who uses AI.

[CryptoHeadlineNews]

At the end of the day I think we'll lose this battle either way. AI bots will get better, humans will get tired.

We are already losing it. ChatGPT is being integrated into our lives and work environments in the same way that the Internet began to be integrated 25 years ago and cell phones 15 years ago. We can swim against the tide all we want in this forum to stay in the dinosaur era as the world continues to progress thanks to the internet, mobile and AI.

I am all for AI, especially if it is used as leverage, where the human rewrites and adapts, and not just copy-pastes.

I totally agree with you on this very statement of yours @ Poker Player, because for me it's never a crime using A.I as an assistant or reference source for creating digital content, only if the user uses it responsibly, but it can only be a crime when such an individual fails to add value and uses the entire generated text completely in its robotic manner. However, A.I is no different from Google where people always go search/source for content when writing projects or creating articles for a blog post, because only if people use it in such a way that adds value or solves a particular life problem. I personally see nothing wrong with using A.I

[dkbit98]

That is the future I fear: the internet as we know it will cease to exist. For now, the AI-spam is still the result of "ask question, copy result, post result", but in a few years it could very well be: "Hey AI, create a bunch of accounts on Bitcointalk, start posting, join signature campaigns, and make me money!", and by the time that works, it's a lost cause for humans.

I am sure there are bigger and more dangerous things someone could do with AI instead of creating accounts and making money from bitcointalk forum.
Creators of AI already started charging for using their services and my predictions is that soon most of the people wont be able to use them anymore, unless they pay good money.
But I agree with you that internet as we know it is living it's last days, maybe reason for introducing internet-kyc for everyone would be exactly this - dangers caused by AI 
This is just my speculation but it's not impossible scenario at all.

[joker_josue]

I am sure there are bigger and more dangerous things someone could do with AI instead of creating accounts and making money from bitcointalk forum.
Creators of AI already started charging for using their services and my predictions is that soon most of the people wont be able to use them anymore, unless they pay good money.

I happen to think the opposite. AI technology is starting to appear in droves. It seems that many of the projects that were just being worked on in the garage are now available to the public. With the mass increase of these platforms, the tendency is for the price to go down. Of course, the best ones will be expensive to use.

I think we're still very much in the phase of being new and then being very popular. Even so, taking into account what the available platforms, in "beta" mode, already do, I believe that this is now always increasing. But we'll have a better idea a year from now.

[karmamiu]

I am sure there are bigger and more dangerous things someone could do with AI instead of creating accounts and making money from bitcointalk forum.
Creators of AI already started charging for using their services and my predictions is that soon most of the people wont be able to use them anymore, unless they pay good money.

I happen to think the opposite. AI technology is starting to appear in droves. It seems that many of the projects that were just being worked on in the garage are now available to the public. With the mass increase of these platforms, the tendency is for the price to go down. Of course, the best ones will be expensive to use.

I think we're still very much in the phase of being new and then being very popular. Even so, taking into account what the available platforms, in "beta" mode, already do, I believe that this is now always increasing. But we'll have a better idea a year from now.

I have to agree on you with that. As of now since AI is still at an early stage, so many are offering this feature for free so that users will have their time to try everything and choose what's best for them, but I think that those projects that are stocked without development before will coming out later this year or early next year. If they couldn't find solution before, maybe now with the aid of AI it will develop faster, and who knows what kind of project they are stocking. Aside from that, there will be tools that can be used freely in public just like how there are options we have now if we want to subscribe or not.

[FatFork]

There are definitely some downsides to AI-generated content that we should be aware of, but I don't think we need to freak out just yet. The internet has already gone through a bunch of changes and I think it'll keep changing in the future. Plus, AI could do some pretty cool stuff for us too, like revolutionize industries and make our lives easier. We just gotta be careful and balance out the good and the bad.

[Out of PATIENCE]

AI scares the bejesus out of me, not only from this forums perspective but throughout the world. The level of sophistication which has already been achieved, should be alarming to everyone. Pondering about how AI could be shifted in biased ways, and be used for malicious purposes, is scary.

Biased adjustments added into an AI's neural network to fit various agendas, manipulate public opinions, controlling herd mentality, and more, could all send the entire world into another level of chaos which we will not understand, until it is too late. Depending on who, or let's be real, which tech giant, corporation, government or special interest group masters it first, likely the latter two.

For this forum, the solution is simple.
AAIM's --- Advanced Artificial Intelligence Moderators.  

Eliminating the need for human moderators all together. Plagiarism will be stifled. But we [disappointingly] will end up with multiple reiterations of the same [hopefully good] content, thru various [sponsored] AI bots, we assume run by humans - but maybe run by other AI's generating and using alts. All information [hopefully accurate] regurgitated in various ways as [maybe real] humans interact with the AI's --- If the content all satisfies the tolerance level of the AAIM's, of course.

Yikes.