• This forum is strictly intended to be used by members of the VS Battles wiki. Please only register if you have an autoconfirmed account there, as otherwise your registration will be rejected. If you have already registered once, do not do so again, and contact Antvasima if you encounter any problems.

    For instructions regarding the exact procedure to sign up to this forum, please click here.
  • We need Patreon donations for this forum to have all of its running costs financially secured.

    Community members who help us out will receive badges that give them several different benefits, including the removal of all advertisements in this forum, but donations from non-members are also extremely appreciated.

    Please click here for further information, or here to directly visit our Patreon donations page.
  • Please click here for information about a large petition to help children in need.

Banning the use of AI and ChatGPT

I do not remember the specifics here well enough to do so currently, but am very open for suggestions. 🙏
 
Reiner04 said:
How about that "a user should post the AI generated arguement or evidences only after cross-checking it with credible sources specific to the topic being discussed" or smth along the line?
Click to expand...
And I think AI can provide sources of what they say so it's always easy to crosscheck and if they can't on something specific then it shouldn't be used as arguement at all.
 
Antvasima said:
I do not remember the specifics here well enough to do so currently, but am very open for suggestions. 🙏
Click to expand...
Possible Suggestion:

The use of AI tools isn't explicitly prohibited to help users formulate their postings on the site (i.e grammar and structure). All threads will be judged by their sources and logic the same as any other. The use of AI tools to provide you sources for an argument is at your own risk as you may be flagged for low-effort postings if your sources do not exist at all.
 
I'm hesitant to vote one way or the other, one feels very cautious in deciding upon something with AI. Much of what the OP says and what they have said since is true, and in particular I wonder if we're recognizing how easy AI makes it to inject false information or to enable stonewalling. As it stands, yes, any person can lie, and any person can attempt to stonewall, but that requires stamina and ideas on their behalf. An AI skips that entirely, and I think many (perhaps including some staff) would be willing to hear out their posts, unknowing of the user's intent, simply because of how good an AI can make it sound. It expands the capability of the user, rarely to our benefit. That the ideas contained may have zero merit but will do everything in their power to convince others that they do, is a concern of mine.

Still, it was aptly pointed out that we would have very little hope of consistently detecting AI. I don't know if we'd have more or less success compared to the academics of the article, but I'm very certain it wouldn't be 100%. Prosecuting a ban on the use of AI, then, is an issue. This isn't strictly an argument against the ban, but it is a troublesome consideration.

Firestorm808 said:
Possible Suggestion:

The use of AI tools isn't explicitly prohibited to help users formulate their postings on the site (i.e grammar and structure). All threads will be judged by their sources and logic the same as any other. The use of AI tools to provide you sources for an argument is at your own risk as you may be flagged for low-effort postings if your sources do not exist at all.
Click to expand...
This is a good start, in my opinion, to creating an AI policy. I think CRTs made with the use of AI should be required to state that they are made with such, even if it is just for translation purposes or what have you, additionally. I figure a staff member would be more alert to AI-related issues if a post is specifically tagged as being AI-generated, otherwise within the rules or not.
 
Thank you for helping out. 🙏

So how about if discussion posts here are only allowed to use artificial intelligence in order to find valid sources of information/reference links or to translate them into proper English, but that our members have to announce that they are doing so in that case? 🙏
 
Antvasima said:
Thank you for helping out. 🙏

So how about if discussion posts here are only allowed to use artificial intelligence in order to find valid sources of information/reference links or to translate them into proper English, but that our members have to announce that they are doing so in that case? 🙏
Click to expand...
Preferably users would then go find the source themselves, to confirm its existence, since AI is known to just. Make them up. But yeah, I'd be fine with that.
 
Thank you for your input help.

Is somebody willing to write up a draft text for rules regarding what we said above then? I can modify the word flow afterwards if necessary. 🙏
 
Antvasima said:
Thank you for your input help.

Is somebody willing to write up a draft text for rules regarding what we said above then? I can modify the word flow afterwards if necessary. 🙏
Click to expand...
I'm not staff, but may I offer to do it, please? What should I type?
 
Firestorm808 said:
Possible Suggestion:

The use of AI tools isn't explicitly prohibited to help users formulate their postings on the site (i.e grammar and structure). All threads will be judged by their sources and logic the same as any other. The use of AI tools to provide you sources for an argument is at your own risk as you may be flagged for low-effort postings if your sources do not exist at all.
Click to expand...
Mr. Bambu said:
I'm hesitant to vote one way or the other, one feels very cautious in deciding upon something with AI. Much of what the OP says and what they have said since is true, and in particular I wonder if we're recognizing how easy AI makes it to inject false information or to enable stonewalling. As it stands, yes, any person can lie, and any person can attempt to stonewall, but that requires stamina and ideas on their behalf. An AI skips that entirely, and I think many (perhaps including some staff) would be willing to hear out their posts, unknowing of the user's intent, simply because of how good an AI can make it sound. It expands the capability of the user, rarely to our benefit. That the ideas contained may have zero merit but will do everything in their power to convince others that they do, is a concern of mine.

Still, it was aptly pointed out that we would have very little hope of consistently detecting AI. I don't know if we'd have more or less success compared to the academics of the article, but I'm very certain it wouldn't be 100%. Prosecuting a ban on the use of AI, then, is an issue. This isn't strictly an argument against the ban, but it is a troublesome consideration.


This is a good start, in my opinion, to creating an AI policy. I think CRTs made with the use of AI should be required to state that they are made with such, even if it is just for translation purposes or what have you, additionally. I figure a staff member would be more alert to AI-related issues if a post is specifically tagged as being AI-generated, otherwise within the rules or not.
Click to expand...
@Omnificence that's what have to be written yet. A draft that constitutes the points and suggestions offered by Bambu and Firestorm here sould be sufficient for now, I believe.
 
Antvasima said:
So how about if discussion posts here are only allowed to use artificial intelligence in order to find valid sources of information/reference links or to translate them into proper English, but that our members have to announce that they are doing so in that case? 🙏
Click to expand...
What about including this part then? 🙏
 
Also, shouldn't more than just warnings and thread bans be on the table for using artificial intelligence to argue for oneself, so we are sufficiently strict to caution our members against doing so? 🙏
 
Antvasima said:
Also, shouldn't more than just warnings and thread bans be on the table for using artificial intelligence to argue for oneself, so we are sufficiently strict to caution our members against doing so? 🙏
Click to expand...
I think it should be handled on a case by case basis. If it is intended to spam or increase word count in an argument to make it look more compelling (as I have seen it in a recent thread where a user did this), I believe a strict warning should be given first as a chance to correct their behavior. If their behavior continues, then we can move to a harsher punishment depending on the situation. But, exceptionally, Considering the severity of the situation, any length of punishment could be justified, including a a normal or permanent ban w/o warning at first.
 
Omnificence said:
Alright, here's what I got. How does it look?
Click to expand...
There is one more thing that needs to be noted, chatgtp translations for source material should be fully disallowed since it will randomly add in parts to the raws/translations without telling the user's. I had a discussion with staff here in the past regarding it, but It was in another thread and in private.

(I received permission to comment regarding this matter in the past, but I don't remember where I wrote it.)
 
Dark_Soul20189 said:
There is one more thing that needs to be noted, chatgtp translations for source material should be fully disallowed since it will randomly add in parts to the raws/translations without telling the user's. I had a discussion with staff here in the past regarding it, but It was in another thread and in private.

(I received permission to comment regarding this matter in the past, but I don't remember where I wrote it.)
Click to expand...
I've already added Ant's suggestion to the policy, and I'm not sure if we should "fully disallow" the use of AI to translate source material. Maybe we could be sure to run the translations through the staff and moderation team if artificial intelligence is ever used for cases like this, but it's either one or the other.
 
Omnificence said:
I've already added Ant's suggestion to the policy, and I'm not sure if we should "fully disallow" the use of AI to translate source material. Maybe we could be sure to run the translations through the staff and moderation team if artificial intelligence is ever used for cases like this, but it's either one or the other.
Click to expand...
The issue is that that doesn't work, since As I have discussed with admins in the past, ChatGPT especially adds texts when translating raws, it doesn't matter if we have the source or the raws if the translations have an additional paragraph added to it when translating it without mentioning it to the users. It makes the entire translation method unreliable.
 
I'm still of the opinion that restricting the use of AI is not beneficial. If we are doing witch hunts based on the suspicion that something is formulated too neatly for a human, then we may as well add the additional similarly hard-to-enforce clause of "and has not been properly checked and thought through by them" or "used it for the purpose of disrupting proper debate".
And I see zero need to declare it then. That's only punishing the honest.

I also generally don't see why AI-generated text shouldn't be on profiles. If a CRT accepts the information provided as accurate, then I don't see why we should reject it. In many cases, it would likely end up actually better written. That seems like a genetic fallacy. The text quality and accuracy of the information is what should matter, not who wrote it.

To make a proposal as to how I would prefer the rules, taking Omnificence's draft and the existing rule as references:
The use of artificial intelligence on the wiki and forum is allowed as long as it is used responsibly and in a not harmful way. To ensure that this is the case, the following rules are in place:
  • When using any AI, such as ChatGPT, it is your responsibility to check the correctness of the information before sharing it with others. This includes verifying that sources and facts obtained through AI are reliable and accurate, any mathematics or formulas used are correct, arguments make logical sense and any generated text accurately conveys the intended meaning. Sharing AI content in a content revision thread or calculation without understanding and fact-checking it can be considered a rule violation.
  • AI-generated images must not be used for profile pictures, verse images, or other official wiki representations unless explicitly approved by staff. The use of AI-generated images in forum threads is allowed, within the same limits as for any human-created picture.
  • AI may be used in discussion threads to assist with gathering ideas for arguments or aid in properly formulating one's own thoughts. Doing both at once, and hence essentially automating the response, should be avoided. Every user has an obligation to ensure that the length of AI-generated texts stay within reasonable limits for the number of arguments provided so that it is not taxing for other involved parties to read the posts.
  • Stonewalling through the use of AI, i.e. to purposefully drag out a debate or to tire out other debaters by posting easy to produce AI generated responses, is prohibited.
Depending on the severity, violations of these rules and guidelines can be punished with warnings, thread bands or bans.

All staff with evaluation rights (Bureaucrats, Administrators, Thread Moderators and, for calculation-related threads, Calculation Group members) have the right to prohibit a specific user from the use of AI within a specific thread if they deem the user's application of it in said thread as harmful. Doing so requires no prior confirmation in the Rule Violation Reports thread. The staff furthermore reserves the right to permanently prohibit specific users to use AI on the wiki and forum, if their use of it is found repeatedly irresponsible or harmful. That decision generally requires prior agreement amongst the staff, for example in the Rule Violation Reports thread.
Click to expand...
Edit: Didn't we have a rule on AI pictures somewhere already or am I misremembering that?
 
Last edited:
DontTalkDT said:
Edit: Didn't we have a rule on AI pictures somewhere already or am I misremembering that?
Click to expand...
I don't think we have rules specific to AI images but I think the rules we have for images in general will generally apply to AI one too. Still I think for AI images, rules gotta be stricter than fanarts.
vsbattles.fandom.com

Editing Rules

The VS Battles Wiki presently consists of 38,036 pages, most of them listing the statistics of characters across different fictional franchises. However, there is a tremendous amount of work involved in maintaining the accuracy of statistics and upholding a high quality for the profiles. In...
vsbattles.fandom.com vsbattles.fandom.com
 
Reiner04 said:
I don't think we have rules specific to AI images but I think the rules we have for images in general will generally apply to AI one too.
vsbattles.fandom.com

Editing Rules

The VS Battles Wiki presently consists of 38,036 pages, most of them listing the statistics of characters across different fictional franchises. However, there is a tremendous amount of work involved in maintaining the accuracy of statistics and upholding a high quality for the profiles. In...
vsbattles.fandom.com vsbattles.fandom.com
Click to expand...
Count me in as vehemently against the use of AI images when it comes to profiles on the site
 
The use of artificial intelligence on the wiki and forum is allowed as long as it is used responsibly and in a not harmful way. To ensure that this is the case, the following rules are in place:

  • When using any AI, such as ChatGPT, it is your responsibility to check the correctness of the information before sharing it with others. This includes verifying that sources and facts obtained through AI are reliable and accurate, any mathematics or formulas used are correct, arguments make logical sense and any generated text accurately conveys the intended meaning. Sharing AI content in a content revision thread or calculation without understanding and fact-checking it can be considered a rule violation.
  • AI-generated images must not be used for profile pictures, verse images, or other official wiki representations unless explicitly approved by staff. The use of AI-generated images in forum threads is allowed, within the same limits as for any human-created picture.
  • AI may be used in discussion threads to assist with gathering ideas for arguments or aid in properly formulating one's own thoughts. Doing both at once, and hence essentially automating the response, should be avoided. Every user has an obligation to ensure that the length of AI-generated texts stay within reasonable limits for the number of arguments provided so that it is not taxing for other involved parties to read the posts.
  • Stonewalling through the use of AI, i.e. to purposefully drag out a debate or to tire out other debaters by posting easy to produce AI generated responses, is prohibited.
Depending on the severity, violations of these rules and guidelines can be punished with warnings, thread bands or bans.

All staff with evaluation rights (Bureaucrats, Administrators, Thread Moderators and, for calculation-related threads, Calculation Group members) have the right to prohibit a specific user from the use of AI within a specific thread if they deem the user's application of it in said thread as harmful. Doing so requires no prior confirmation in the Rule Violation Reports thread. The staff furthermore reserves the right to permanently prohibit specific users to use AI on the wiki and forum, if their use of it is found repeatedly irresponsible or harmful. That decision generally requires prior agreement amongst the staff, for example in the Rule Violation Reports thread.
Click to expand...
  • "Approved by staff" can mean quite a lot and I think it's a poor phrasing. Can an Image Helper approve such an image? Can a calc group member? Translation helper? Is it restricted to administrators and above? Is it the same line as your later bit about restricting specific users from using AI in a thread? This needs clarification.
  • "should be avoided" does not render it against the rules, although given its presence in the rules, I find this vague. Do you propose we punish users if they do both? Is this not equally as difficult to determine as other AI-related offenses, which was your reasoning as to why we should not bother with them? Your fourth bullet point also seems to relate to this.
  • Furthermore, if we already insist that users adhere to rules regarding AI, then why not add the notification that a post used AI in its creation? I fail to see how it is "punishing the honest", if we'd consider it a rule violation to lack that. It would alert staff members to approach these threads with greater scrutiny- I genuinely do not see a downside.
 
Mr. Bambu said:
  • "Approved by staff" can mean quite a lot and I think it's a poor phrasing. Can an Image Helper approve such an image? Can a calc group member? Translation helper? Is it restricted to administrators and above? Is it the same line as your later bit about restricting specific users from using AI in a thread? This needs clarification.
Click to expand...
I would suggest to have this operate under the same standards as CRT approval. So perhaps instead of "unless explicitly approved by staff" we can just say "unless approved in a content revision thread".
Mr. Bambu said:
  • "should be avoided" does not render it against the rules, although given its presence in the rules, I find this vague. Do you propose we punish users if they do both? Is this not equally as difficult to determine as other AI-related offenses, which was your reasoning as to why we should not bother with them? Your fourth bullet point also seems to relate to this.
Click to expand...
Personally, I see this as something to be discouraged, but not outright to be banned. Hence the formulation.
That's a part I am willing to compromise on, though.

And my point regarding AI-related offences being difficult to determine wasn't quite that we shouldn't do it. It was that either we don't try it or, if we do, then we can just as well afford to actually distinguish between good and bad usage, as that isn't much harder either at that point.
The suggested ruleset goes more in the latter direction for the sake of finding common ground.
Mr. Bambu said:
  • Furthermore, if we already insist that users adhere to rules regarding AI, then why not add the notification that a post used AI in its creation? I fail to see how it is "punishing the honest", if we'd consider it a rule violation to lack that. It would alert staff members to approach these threads with greater scrutiny- I genuinely do not see a downside.
Click to expand...
The downside is that, as virtually every place on the internet currently shows, here included, many people are extremely willing to engage in genetic fallacy when it comes to any content touched by AI.
Declared AI usage, even if within the rules, is likely to be criticized frequently when it happens (unless we make a rule to ban that) and even if not, it certainly will make the actual points made be viewed more negatively.

Those who followed that rule would have a tangible downside. Those that would follow the rule are also not those I believe would have trouble following the ruleset. The people with whom we have problems would primarily be those who wouldn't declare it anyway.
 
We are a human institution, and thus we are limited by human flaws. It is practically every day that some mistake is found in how we've done something. It's to be expected. Inspiring bias or not, I see no reason to not tag posts as a rule. If we, the staff, are conscious that what we're reading was at least partly AI generated, then we can be more careful in accepting it. We make mistakes too frequently as it is. I cannot agree to the lack of such a rule.
 
I personally largely agree with Bambu here.
Mr. Bambu said:
  • "Approved by staff" can mean quite a lot and I think it's a poor phrasing. Can an Image Helper approve such an image? Can a calc group member? Translation helper? Is it restricted to administrators and above? Is it the same line as your later bit about restricting specific users from using AI in a thread? This needs clarification.
Click to expand...
Perhaps administrators and above, to stay on the safe side?
Mr. Bambu said:
  • "should be avoided" does not render it against the rules, although given its presence in the rules, I find this vague. Do you propose we punish users if they do both? Is this not equally as difficult to determine as other AI-related offenses, which was your reasoning as to why we should not bother with them? Your fourth bullet point also seems to relate to this.
Click to expand...
I agree about that "should be avoided" should be changed to "is not allowed". I do not want to exhaust our staff members by having them be forced to argue against tireless machines.
Mr. Bambu said:
  • Furthermore, if we already insist that users adhere to rules regarding AI, then why not add the notification that a post used AI in its creation? I fail to see how it is "punishing the honest", if we'd consider it a rule violation to lack that. It would alert staff members to approach these threads with greater scrutiny- I genuinely do not see a downside.
Click to expand...
I also agree about that we should demand full disclosure about using a.i. for assistance.

In addition, I agree with DontTalk keeping the wording that all of the data found through the use of a.i. needs to be verified by oneself afterwards, and I also definitely do not support using it for translations, as it may get "creative" and fill in the blanks or adjust the wording or context in that regard. 🙏
 
- I think a CRT for AI generated art is a bit much. I'm content with restricting it to administrators but I'd agree with letting it be moderators, too- even content mods. I don't think it's an issue worth cracking down too hard on. AI images are controversial but really we can just say "no" if there's no good reason to use one. I would trust our staff to do that.

- Broadly speaking, aside from what protest I've already made, I accept DT's wording as well.
 
DontTalkDT said:
I would suggest to have this operate under the same standards as CRT approval. So perhaps instead of "unless explicitly approved by staff" we can just say "unless approved in a content revision thread".

Personally, I see this as something to be discouraged, but not outright to be banned. Hence the formulation.
That's a part I am willing to compromise on, though.

And my point regarding AI-related offences being difficult to determine wasn't quite that we shouldn't do it. It was that either we don't try it or, if we do, then we can just as well afford to actually distinguish between good and bad usage, as that isn't much harder either at that point.
The suggested ruleset goes more in the latter direction for the sake of finding common ground.

The downside is that, as virtually every place on the internet currently shows, here included, many people are extremely willing to engage in genetic fallacy when it comes to any content touched by AI.
Declared AI usage, even if within the rules, is likely to be criticized frequently when it happens (unless we make a rule to ban that) and even if not, it certainly will make the actual points made be viewed more negatively.

Those who followed that rule would have a tangible downside. Those that would follow the rule are also not those I believe would have trouble following the ruleset. The people with whom we have problems would primarily be those who wouldn't declare it anyway.
Click to expand...
Firestorm808 said:
Personally, all I care about is accuracy and sincerity. If AI generated text is straight to the point and verified as accurate, what's the problem?
Click to expand...
Mr. Bambu said:
- I think a CRT for AI generated art is a bit much. I'm content with restricting it to administrators but I'd agree with letting it be moderators, too- even content mods. I don't think it's an issue worth cracking down too hard on. AI images are controversial but really we can just say "no" if there's no good reason to use one. I would trust our staff to do that.

- Broadly speaking, aside from what protest I've already made, I accept DT's wording as well.
Click to expand...
Alright, here's the updated version.
vsbattles.fandom.com

Regarding ChatGPT and other uses of AI

vsbattles.fandom.com vsbattles.fandom.com
 
Mr. Bambu said:
We are a human institution, and thus we are limited by human flaws. It is practically every day that some mistake is found in how we've done something. It's to be expected. Inspiring bias or not, I see no reason to not tag posts as a rule. If we, the staff, are conscious that what we're reading was at least partly AI generated, then we can be more careful in accepting it. We make mistakes too frequently as it is. I cannot agree to the lack of such a rule.
Click to expand...
AI generated texts are not magically more convincing. If they are more convincing at all, it is at best because they lay out the arguments well which is a desirable trait, given that we wish to have an easy time understanding the points made.
Being extra cautious because it is AI is straight up an undesirable thing in my books, since it means you treat the arguments differently depending on the source. I.e. you suggest we should actively commit a genetic fallacy.
All arguments should be evaluated with the same scrutiny regardless of source. There should be no "being more careful in accepting it" due to who wrote it.

If anything your argument makes me more adamant in not having that rule, as it shows my worries are justified. You actively wish to ensure that posts involving AI are less likely to be accepted.
 
DontTalkDT said:
AI generated texts are not magically more convincing. If they are more convincing at all, it is at best because they lay out the arguments well which is a desirable trait, given that we wish to have an easy time understanding the points made.
Being extra cautious because it is AI is straight up an undesirable thing in my books, since it means you treat the arguments differently depending on the source. I.e. you suggest we should actively commit a genetic fallacy.
All arguments should be evaluated with the same scrutiny regardless of source. There should be no "being more careful in accepting it" due to who wrote it.

If anything your argument makes me more adamant in not having that rule, as it shows my worries are justified. You actively wish to ensure that posts involving AI are less likely to be accepted.
Click to expand...
You should treat arguments differently based on the source. There are more reliable sources than others. This is the basics. If you have Source 1 ("my aunt kelly said so") vs Source 2 ("Harvard University put out a paper stating that this was the case") and you say "The sources are equally valid" then you aren't paying attention. Genetic fallacy only applies when there is actually no reason to distinguish between one source or the other. In this case, we have one! AI lies. A lot. Giving the heads up that something was written by it, if only in part, is solely beneficial.

And you're right, all arguments should be treated with the same level of scrutiny- but they aren't, because we are humans and we make human errors. And... what wrote it, not who.

Your position is based in discrimination against a machine which consistently generates lies. That has been laid out across this thread and it would be very easy to get particular examples abroad. I am similarly adamant that if these are the counterarguments to such a rule that I cannot, in good conscious, allow the rule to not exist.
 
Mr. Bambu said:
You should treat arguments differently based on the source. There are more reliable sources than others. This is the basics. If you have Source 1 ("my aunt kelly said so") vs Source 2 ("Harvard University put out a paper stating that this was the case") and you say "The sources are equally valid" then you aren't paying attention. Genetic fallacy only applies when there is actually no reason to distinguish between one source or the other. In this case, we have one! AI lies. A lot. Giving the heads up that something was written by it, if only in part, is solely beneficial.
Click to expand...
I think they mean that the initial claim (AI or not) and the actual sources for the claim are two different things.

For example, the AI might say that Thor did X feat in Issue Y. In that same vein, a random user can say that Thor did X feat in Issue Y.

In either case, a claim is made, but the actual scans and source links still need to be provided to support said claim.

AI and a random user can claim whatever they want. At the end of the day, it's the staff that are verifying the sources provided for that claim.
 
You must log in or register to reply here.

Similar threads

Kisaragi_Megumi
  • Locked
Are AI really allowed on the forum?
Replies
6
Views
916
Vietthai96
Vietthai96
Fan Calculation: AI-2027 and Orion's Arm Intelligence Scale Comparison
Replies
0
Views
561
User_A_98732165489
Meet The Consensus-1 (AI-2027)
Replies
3
Views
1K
Oiguana2701
Oiguana2701
Edutyn
What would this give?
Replies
1
Views
440
DenzOnigiri
DenzOnigiri
AppleMaker
Can I make a profile of stereotypical Twitter users?
Replies
3
Views
765
AppleMaker
AppleMaker
Back
Top