This forum will need explicit rules on the use of AI.

[digaran]

Question is, where does the AI get it's ideas/ content from?
If we see a someone praising bitcoin and LN to later find out it was AI, who's ideas was posted?

A forum's community needs genuine discussions, what if I used the AI but didn't really agreed with the results but posted it anyways to eaither. e.g., farm merits or earn easy sig money? the spirit of humanity will be lost if that happens.

[1715086814]

1715086814

[1715086814]

1715086814

[1715086814]

1715086814

[1715086814]

1715086814

[NotATether]

If an AI is able to consistently create content that doesn't break any of the forum's rules (good quality, on-topic, not just a padded word salad, no plagiarism, etc, etc), I, for one, welcome our new machine overlords. Otherwise, content that violates the rules, AI or human produced, can already be dealt with with our current rules and policies in mind.

Alright. I'll see if I can contract Aaron Schwarzenegger to build me one last Terminator for this purpose.

On a more serious note, campaign managers should exclude AI-generated posts from the payroll as soon as possible, otherwise this feature will be abused as ChatGPT sometimes makes incorrect answers.

Campaign managers: Use this tool to detect AI-generated text: https://huggingface.co/openai-detector. I recommend considering anything with a score of 90%+ as AI-generated.

[Don Pedro Dinero]

On a more serious note, campaign managers should exclude AI-generated posts from the payroll as soon as possible, otherwise this feature will be abused as ChatGPT sometimes makes incorrect answers.

Campaign managers: Use this tool to detect AI-generated text: https://huggingface.co/openai-detector

I am sure good campaign managers who care about quality will, but I am not so sure about campaigns like 1xBit and maybe some others.

[Solosanz]

Campaign managers: Use this tool to detect AI-generated text: https://huggingface.co/openai-detector. I recommend considering anything with a score of 90%+ as AI-generated.

It's useful, but it's not really a perfect tool to detect AI generated text, see below:

First I got result 0.80% real.

It is theoretically possible for the price of bitcoin to drop to zero, but it is highly unlikely to happen. There are several reasons for this.

First, bitcoin has a fixed supply, which means that there is a limited amount of bitcoins that can be mined. This scarcity gives bitcoin value and makes it unlikely that the price will drop to zero.

Second, bitcoin has a decentralized network of users who are committed to the success of the currency. These users are motivated by a belief in the value of bitcoin and the underlying technology (blockchain) and are likely to continue using and holding bitcoin even if the price drops temporarily.

Finally, bitcoin has a strong track record of price stability over its 12+ year history, despite experiencing some significant price fluctuations. While it is always possible for the price of bitcoin to go down, it is unlikely that it would drop to zero.

After some edit on punctuation, I got result 99.58% real.

It is theoretically possible for the price of bitcoin to drop to zero, but it is highly unlikely to happen. There are several reasons for this

First,, bitcoin has a fixed supply, which means that there is a limited amount of bitcoins that can be mined. This scarcity gives bitcoin value and makes it unlikely that the price will drop to zero.

Second,, bitcoin has a decentralized network of users who are committed to the success of the currency. These users are motivated by a belief in the value of bitcoin and the underlying technology (blockchain) and are likely to continue using and holding bitcoin even if the price drops temporarily.

Finally,, bitcoin has a strong track record of price stability over its 12+ year history, despite experiencing some significant price fluctuations. While it is always possible for the price of bitcoin to go down, it is unlikely that it would drop to zero

Actually AI generated text are giving a systematic explanation and mostly about common sense, this could be a way for the managers to spot AI generated text without using a tools.

I am sure good campaign managers who care about quality will, but I am not so sure about campaigns like 1xBit and maybe some others.

Why we need to care with 1xbit scam manager and other managers that get payment with worthless tokens? they're a rubbish in this forum and there's no way to control them except create self moderated threads to delete their low quality post or just report to moderators.

[NotATether]

After some edit on punctuation, I got result 99.58% real.

Substantial edits to AI generated text can no longer be classified as such, but as written by a human. Thus, managers should choose a confidence score that will withstand "lazy adjustments" to punctuation and typos to try to drop the score down.

Of course, if you see a post full of typos, then it's not AI generated and in most campaigns the person writing such posts would be considered incompetent and thrown out.

[Stalker22]

Question is, where does the AI get it's ideas/ content from?
If we see a someone praising bitcoin and LN to later find out it was AI, who's ideas was posted?

A forum's community needs genuine discussions, what if I used the AI but didn't really agreed with the results but posted it anyways to eaither. e.g., farm merits or earn easy sig money? the spirit of humanity will be lost if that happens.

If people are just pretending to agree or disagree with something they read by copying the replies of an AI, they are basically just faking it until they make it. And that is not exactly the best way to build a genuine community, is it? It is like trying to win a game of poker with a deck of all jokers. Sure, you might get lucky and fool someone for a while, but eventually, your bluff will be called. So let us all agree to just be real with each other and let us celebrate and embrace our individual differences, and create a space where we can all feel valued and special, like a royal flush!

[Ultegra134]

If an AI is able to consistently create content that doesn't break any of the forum's rules (good quality, on-topic, not just a padded word salad, no plagiarism, etc, etc), I, for one, welcome our new machine overlords. Otherwise, content that violates the rules, AI or human produced, can already be dealt with with our current rules and policies in mind.

I agree. Whether we like it or not, AI content is not necessarily plagiarized. As long as we don't have duplicate content, which might occur if a large number of users use a specific platform on a specific question such as "Should I invest in Bitcoin?" it doesn't do much harm on the forum if used in moderation. Imagine 90% of users being practically bots; that would be disastrous.

Campaign managers should prohibit the usage of AI, which would actually discourage users from using AI in their posts.

[Igebotz]

If an AI is able to consistently create content that doesn't break any of the forum's rules (good quality, on-topic, not just a padded word salad, no plagiarism, etc, etc), I, for one, welcome our new machine overlords. Otherwise, content that violates the rules, AI or human produced, can already be dealt with with our current rules and policies in mind.

Alright. I'll see if I can contract Aaron Schwarzenegger to build me one last Terminator for this purpose.

On a more serious note, campaign managers should exclude AI-generated posts from the payroll as soon as possible, otherwise this feature will be abused as ChatGPT sometimes makes incorrect answers.

Campaign managers: Use this tool to detect AI-generated text: https://huggingface.co/openai-detector. I recommend considering anything with a score of 90%+ as AI-generated.

The GPT-2 you provided only works with ChatGPT generated posts and the site prediction is inaccurate, I tried the site twice with original local Naija posts and received two different predictions, ChatGPT is not the only AI-generated platform that people use to post on forums. It is unreliable to use https://huggingface.co/openai-detector. We are only going to force traffic on the site to make them more money nothing more.

[mindrust]

That's all. Surely the owner and the people most involved in it are thinking about it, but AIs are going to be used on the forum if they are not already being used, as they are being used all over the internet.

Better to come up with clear rules than to go willy-nilly.

How are you going to identify if the user is using AI in his posts? How is anybody going to do that? Do people even have that much free time?

I think whether it is against the rules or nor won't change a thing. You can even command AI to create unique posts.

You just tell it to write something about something, give the post length etc and it is going to make you one.

AI will own everybody.

[digaran]

How are you going to identify if the user is using AI in his posts?

I will not disclose any accurate source for my claim but there is actually no way to find that out if someone is using AI, since there are services out there for free making and editing your AI generated content in a way to avoid 99+% detection.
IMO anyone caught using AI generated content as their own should be banned, I wouldn't want to engage in a discussion where the other person is not really a person.

We want to engage and discuss with real people with their own real life achievements. Just as an example, I would like to read the posts of gmaxwell to know him better and more, just imagine if I wanted to follow him as a role-model and later find out or not to find out he was actually an AI behind the account.

What I don't like about this subject is the fact that AI could greatly influence us in the long term without us realizing it.

[Accardo]

How are you going to identify if the user is using AI in his posts?

I will not disclose any accurate source for my claim but there is actually no way to find that out if someone is using AI, since there are services out there for free making and editing your AI generated content in a way to avoid 99+% detection.
IMO anyone caught using AI generated content as their own should be banned, I wouldn't want to engage in a discussion where the other person is not really a person.

We want to engage and discuss with real people with their own real life achievements. Just as an example, I would like to read the posts of gmaxwell to know him better and more, just imagine if I wanted to follow him as a role-model and later find out or not to find out he was actually an AI behind the account.

What I don't like about this subject is the fact that AI could greatly influence us in the long term without us realizing it.

The same way you wouldn't like to communicate with a person that doesn't write correct grammar or keeps repeating phrases in their content is how AI writing works, users vary. If a person is skilled; knows how to utilize AI like chatGPT their response will be as good as that of a good writer. Additionally, the app can be asked to insert synonyms to a keyword to avoid repetitions. Also, every technology offers a new opportunity for users on the internet, in future skilled Chatgpt users can get hired to do a great writing job including editing.

Moreover, you are not communicating with a bot, as an AI doesn't have a brain it's controlled by humans, it just hurts that the next person is using a third-party app to communicate with you making it easier for him to write, while you chat genuinely with your brain. However, it's sad that AI is hitting hard on the internet, and mainly focusing the heat on writing communities, but in the long run, nothing can be done about it, imagine ChatGPT has over a million users within a short period. It's a technology that needs adoption, and writers don't seem to like the idea, as if it'll take their jobs from them, which is not true, contrarily it'll make their jobs easier.

In a forum like this, it'll be hard not to detect AI-generated contents if the user didn't amend it, but if the changes are correctly done Chatgbt can be said to be used for research purposes - easing the research process for a user. If he edits and publishes the thread, it's not completely AI-generated. You can say to be communicating with a bot if the user published the generated post without proofreading it. Moreover, those answers provided by chatGPT are generated from google, and plagiarism detection all depends on the user and the kind of query they gave to the bot. Like the moderator said; if the user is skilled with the use of AI that it doesn't break any rules then it is not a problem in any form.

I'd suggest that an AI detector is added to the forum to help users crosscheck if a thread is poorly generated by an AI, such a thread can be reported to mods pointing out how it broke a rule in the forum. It'll ease the stress for moderators and keep the community as clean as usual. Finally, what need is it to pay to write in the forum? soon these AI writing apps will get expensive to use regarding the speed people are signing up to utilize them.

[krishnaverma]

I also think that forums will be affected by growing popularity of AI and chatgpt. Especially a forum like bitcointalk where users are rewarded for posts if they are in signature campaign. Some people will try to game the system. There can be arrangements made to combat this but it will be huge load on forum staff and resources.

[PawGo]

I also think that forums will be affected by growing popularity of AI and chatgpt. Especially a forum like bitcointalk where users are rewarded for posts if they are in signature campaign. Some people will try to game the system. There can be arrangements made to combat this but it will be huge load on forum staff and resources.

Can't wait when purely AI account reach high Position (rank) and would be accepted into a signature campaign. It would be one small step for an account, one giant leap for AI.

And seriously, we should start thinking about it as about realistic scenario - 10 years ago the discussion we have now have been seen as an abstract academic dispute without any relation to real life.

[Poker Player]

I also think that forums will be affected by growing popularity of AI and chatgpt. Especially a forum like bitcointalk where users are rewarded for posts if they are in signature campaign. Some people will try to game the system. There can be arrangements made to combat this but it will be huge load on forum staff and resources.

Can't wait when purely AI account reach high Position (rank) and would be accepted into a signature campaign. It would be one small step for an account, one giant leap for AI.

And seriously, we should start thinking about it as about realistic scenario - 10 years ago the discussion we have now have been seen as an abstract academic dispute without any relation to real life.

I agree with this, and expressed it in another thread, but with a nuance. It is not simply an "AI account", it would be an account created by a Bitcointalk member who uses the AI but tweaks the texts that the AI writes, adapting them accordingly. In other words, it would be using the AI's leverage. It would be a symbiosis between the Bitcointalk member and the IA, so that a new account can be created that writes faster and at least in some cases better.

If it is seen to work, this could result in being able to manage maybe 4 accounts instead of 2 as many have and get paid for them. Obviously, without being a moron by exchanging merits between the accounts or joining the same campaign, etc.

This is all hypothesis and I am not sure if this could be really the case. I would like to try it but I do too many things already.

I would bet that there is currently someone in the top 50 of merit earned who is testing this.

[NotATether]

If it is seen to work, this could result in being able to manage maybe 4 accounts instead of 2 as many have and get paid for them. Obviously, without being a moron by exchanging merits between the accounts or joining the same campaign, etc.

That won't work. All those accounts will get the same fingerprint which will make it easy for anti-multiaccounters to catch and report them. AI is not known to write uniquely (I mean, you can get it to write in some weird tone but at that point, everyone will know it's a bot).

I suggest a form of text CAPTCHA where a certain type of question is asked that only humans know the answer to. For example: "why did dasher lose his funds?"

[Poker Player]

That won't work.

Have you even read what I wrote? It seems you have been busy again.

All those accounts will get the same fingerprint which will make it easy for anti-multiaccounters to catch and report them.

To report them for what? Having multiple accounts is not forbidden.

AI is not known to write uniquely (I mean, you can get it to write in some weird tone but at that point, everyone will know it's a bot).

I suggest a form of text CAPTCHA where a certain type of question is asked that only humans know the answer to. For example: "why did dasher lose his funds?"

Read again, and please next time you quote me try to (agree or) refute what I say. Not to apparently refute what I haven't said.

[Plaguedeath]

All those accounts will get the same fingerprint which will make it easy for anti-multiaccounters to catch and report them.

To report them for what? Having multiple accounts is not forbidden.

Yep, this is why the @OP creating this thread.

However I think it will depends on the managers and merit sources, merit sources need to spend more time to look the post history before merit them and managers need to pay attention to read the quality post and not only from the merit they got in last 120 days.

But don't expect anything if you're talk about bounty hunters, it's already cursed lol.

[Findingnemo]

The growth of AI is really getting out of the hand and already one of the forum user highlighted the reply created by bot when the question has been asked and its not that bad and it doesn't create any suspicion until we spend time and detect the actual source of the content because what AI doing is simply gathering information all over the internet and try to provide a reasonable answer.

AI should be treated as bot account if someone is really willing to use bot account for any such purpose and it should be exempted from signatures so abuser won't get benefits of signature campaigns.

[digaran]

The use of AI to generate material for forums is highly questionable and unethical. Instead of putting in the necessary effort and creativity, some people rely on artificial intelligence to do the work for them, resulting in a lack of originality and authenticity.

This is not only lazy and unoriginal, but it also constitutes stealing someone else's ideas or work and passing it off as your own. This is a serious breach of ethics and integrity that should not be condoned or encouraged in any way. Furthermore, using AI to generate material can be misleading and confusing for readers. It is essential for forum participants to know the source and authenticity of the information they are consuming, and AI-generated material does not provide this transparency.

As an AI hater, I strongly discourage the use of artificial intelligence to generate material for forums. It undermines the integrity of the forum and the value of original, authentic material. We should prioritize originality and effort in our online communication and not rely on shortcuts such as AI-generated material.

😅🤣





In red is what I did, it's fun to use it against itself

[NotATether]

That won't work.

Have you even read what I wrote? It seems you have been busy again.

All those accounts will get the same fingerprint which will make it easy for anti-multiaccounters to catch and report them.

To report them for what? Having multiple accounts is not forbidden.

AI is not known to write uniquely (I mean, you can get it to write in some weird tone but at that point, everyone will know it's a bot).

I suggest a form of text CAPTCHA where a certain type of question is asked that only humans know the answer to. For example: "why did dasher lose his funds?"

Read again, and please next time you quote me try to (agree or) refute what I say. Not to apparently refute what I haven't said.

Basically I was trying to tell you that the monotony of AI systems like ChatGPT will make it easy to spot all the people writing with AIs because they will sound almost exactly the same. And the effort required to rewrite the text to make it unique for each user will be so tedious that people will either abandon this whole enterprise or just copy-paste AI text regardless and face the consequences relevant  for their particular circumstances-e.g. if they are sig campaign members they are all thrown out.

I have no problem with AI accounts used for information or humor purposes, I actually made one myself on the Off Topic board.