Policy Update A Final Word on AI In Roleplay

[JamesMartin]

Evening Sanctum.

This is, frankly, not the announcement I wanted to make tonight. We've been working on some changes today that I think will help people find partners more easily and with more nuance.

However, that is ill-suited to the tenor of the rest of this announcement.

In short, I was made aware today of at least two members who we directly confirmed were using AI to write large sections of their roleplay posts.

I will not name them, anyone who has done this knows who they are and it is possible that our past mentions of the topic have been ambiguous. I wish to remove any ambiguity:

Any user who we find to have written any substantial portion of a roleplay post with AI will be immediately and permanently banned.

They will also be reported to the admins of all other 18+ sites (yes, we all talk behind your backs.)

To put this in more detailed terms:

1. We define "substantial portion" as 20% of the reply. This is to allow exceptions for genuine uses like grammar checking.

2. Use of AI in the process of writing is allowed. This applies only to the words that appear in a final post, we don't mind if you are asking chat GPT "what's another word for flaccid" or to explain the Roman Senate. As long as you are using what you get to write instead of copy-pasting it, you are in the clear

3. This will exclude any use of AI which is deliberately subversive (like using AI to write for a computer character), so long as this is clear.

4. This policy does not apply to images unless you directly claim they are your own work

5. If you have any questions, please ask me directly, either in this thread in in PM.

We consider this a matter of user respect. Writers on this site put an immense amount of time, effort and love into their posts. To respond to that with copies of the regurgitation of a robot is an unacceptable breach of trust.

I will add that we are aware of the drawbacks of online AI checkers and will not be unthinkingly relying on them. We have however tested ways that we feel offset the risk of a false positive. We have also found methods of detection that do not rely on online checkers.

As of the time this message posts, we will be enacting a one-week amnesty to allow for this message to be understood crystal clear. Anyone who violates this rule after that point will be banned with no further warning. Our rules have been updated with this policy and we will strictly enforce it.

Additionally

This conversation brought up a longstanding issue we have grappled with as staff: How do we handle alerting users that their partners are banned?

We have in the past privately messaged users, but this is imprecise (it can be hard to find everyone who needs to know) and time-consuming.

We have considered a public ban forum, but that feels like it will only cultivate drama.

The compromise we have come to is this:



This extremely ugly username and banner will be applied to any banned user who we consider "established". In other words, anyone who has multiple active roleplays on the forums (underage users banned instantly, trolls, etc, will be skipped) will have this applied in order to inform any partners of what has happened.

This will not apply to deleted users.

As the banner says, anyone seeking clarification for a ban reason is free to contact an admin. We might not disclose full details, but we will tell you what we can.

I want to be clear, this is a niche situation. I can count on one hand the number of established users we ban in a given year. It is, however, something that is sometimes necessary.

I sincerely hope that no one reading this is going to give me cause.

 

[Blackstone]

A question about AI Art? We are allowed to use it as long as we clearly state it is AI, it is not our own creation, and only using it to bring to life the concept of a scene or a character. I have been over the pros and cons of AI art and I empathize with the cons of it and the arguments of content creators, at the same time I am not flush with Scooby Snacks to shill out for commissions for original work. I mean my stick figers and caveman like sketches does not quite get across what I am trying to describe.

 

[JamesMartin]

We are allowed to use it as long as we clearly state it is AI

Our rules on AI art are:

1. It does not go in the art forum

2. You cannot claim it is your own work.

As long as neither of those is violated, you are in the clear.

 

[NotAllThatCreative]

Thank you all for your hard work! You guys are the best

 

[JamesMartin]

To answer some questions publically I have received in PMs:

1. No, using Grammarly is not an issue. Grammarly does not use AI and even if for some reason it did in the future, our concern is with AI doing the writing for you, not as a tool to make your writing better

2. Writing in documents off-site does not cause issues, nothing about copypasting your own work will make it resemble AI

3. Our rules for images are far laxer, as we are not an art site. AI images are fine, as long as they are not posted in the art forum and you do not claim they are your own artwork.

 

[Summer Rain]

Great clarification! I cannot help but wonder why someone would use AI to "write" a post. Why not just read a book?!

 

[DaylitDreamR]

When a user is suspected of using AI to misrepresent/forge their writing abilities, will there be an opportunity to speak with the investigating admin(s), in defence, before a final decision is rendered and the account is banned??

I support blacklisting people that abuse the use of AI technology or engage in plagiarism. But due to the fact these bans will have consequences that reach beyond our community, to other Roleplays communities, for the individual(s) in question. I just want to make sure there are adequate checks and balances in place, because it is not a light accusation to have leveled against you.

Damn, I kinda sound like a lawyer...

 

[Chernabog]

I'm also in agreement with the post above/mutually curious, to wonder what the checks and balances are coming with this adjustment. Already we're seeing the unintended repercussions of actual artists having their art banned/suspected of being AI when no, they drew it themselves and unfortunately AI has scraped their art and it takes a lot of effort to get themselves unbanned and also having to show their drawing process.

I ask specifically in the case of writing because for example, I recently posted some of my personal writing on a site and discovered that apparently a lot of people suspect anything 'purple-prose' or adjacent to be AI written because of its complexity? I'm concerned that having flowery language/etc is gonna trigger something here and then others or myself have no warning/way to prove otherwise.

 

[Brigand]

I don't understand why anybody would ever get an A.I. to write for them. Isn't the point of writing to learn and study how to write? To appreciate the craft/art/nuance that goes into sentence construction? I don't get it.

That said, I hope due diligence is done in reporting/banning people for being suspected of using A.I. -- Imo, this is slightly abusable. All someone needs to do to throw a spanner in the works if they dislike another user is to claim they use A.I. to write their content. Maybe just a heads up. It seems hard to prove in my opinion, especially if someone is ''splitting'' (using A.I. to write some of their work, then rewriting it in their own prose). Would be pretty grey if someone was to come along and abuse this rule like that, but I can see it happening, personally.

 

[Nobilis]

I ask specifically in the case of writing because for example, I recently posted some of my personal writing on a site and discovered that apparently a lot of people suspect anything 'purple-prose' or adjacent to be AI written because of its complexity? I'm concerned that having flowery language/etc is gonna trigger something here and then others or myself have no warning/way to prove otherwise.

I have never seen purple prose to meaningfully matter in these questions.

 

[Miree]

I wonder about AI text checking/proofreading - since English is not my first or even second language, I very often run posts through AI to correct not only grammar, but also stylistic errors and make sentences more natural in English. Is it abuse to put the post into AI and post with stylistic corrections and improved flow? I've always treated it as help with language issues and simply learning how to write in foreign language, so that language barriers do not take away my pleasure in role-playing and creating, but I would not like to trigger a ban, and even more so, offend someone by using it this way

I would like to emphasize that the scenes are entirely created and written by me, I am only concerned about the stylistic issue and make sure that such use is permitted

 

[JamesMartin]

... one time I fall asleep at a sane hour all the good questions get posted

When a user is suspected of using AI to misrepresent/forge their writing abilities, will there be an opportunity to speak with the investigating admin(s), in defence, before a final decision is rendered and the account is banned??

I support blacklisting people that abuse the use of AI technology or engage in plagiarism. But due to the fact these bans will have consequences that reach beyond our community, to other Roleplays communities, for the individual(s) in question. I just want to make sure there are adequate checks and balances in place, because it is not a light accusation to have leveled against you.

Outside of users we confirm are underage or ones who engage in open harassment, there is always a process before a ban. We're not going to accuse anyone of doing this lightly and we're going to speak to that person first.

I ask specifically in the case of writing because for example, I recently posted some of my personal writing on a site and discovered that apparently a lot of people suspect anything 'purple-prose' or adjacent to be AI written because of its complexity? I'm concerned that having flowery language/etc is gonna trigger something here and then others or myself have no warning/way to prove otherwise.

So, this is a hard question to give specifics on because, if you explain the ways you have found that give away something is written by AI, you tell the people who use it what to fix.

Suffice it to say: We're not relying on purple prose as an indicator something was written by AI. At all. The indicators we have found are all mistakes that are common for AI but no human writer would make.

All someone needs to do to throw a spanner in the works if they dislike another user is to claim they use A.I. to write their content.

We're cognizant of that and we are definitely not banning based on accusation alone. A claim like that would be investigated, but it would be staff who decide what the content is.

I wonder about AI text checking/proofreading - since English is not my first or even second language, I very often run posts through AI to correct not only grammar, but also stylistic errors and make sentences more natural in English. Is it abuse to put the post into AI and post with stylistic corrections and improved flow? I've always treated it as help with language issues and simply learning how to write in foreign language, so that language barriers do not take away my pleasure in role-playing and creating, but I would not like to trigger a ban, and even more so, offend someone by using it this way

This would not result in a ban. Our core issue is with situations where the person is using the AI to skip the writing process. Using it to check something you have already written is, to my mind, no different than using a spelling or grammar checker, which is something I suspect most of us do.

 

[Lyndis]

This is such a mind-boggling thing to read. Part of the joy in RP (for myself, anyway) is the act of writing and creating. Thinking of clever ways to twist the story, or how a particular character will react to an event, the excitement of knowing the response I'm about to send is going to make my partner seethe/stress/sigh. What engagement or thrill is there to be had when you've essentially put that task to auto-pilot?

Personal Anecdote Time:
My brother and I used to play World of Warcraft during it's heyday and said younger brother ended up having two of his accounts banned because he paid a third party to gold farm and powerlevel his characters to the maximum level cap. I was utterly confused on why he did this. The whole purpose of the game is to play through it. Ostensibly, you play games for enjoyment. The experience of it lives within the quests and dungeons that he paid someone else to finish for him. All to get him as quickly as possible to max level where he could... what?

What was the end goal? He wouldn't have gained the knowledge from previous dungeons to properly face the bosses waiting in the end game instances. He wouldn't even know how to properly run his own character, as his character's skill tree would have been specc'd by someone else and had abilities he'd never used before at lower levels. All he had was the end result. A max level character, with end game armor, and with an end game mount to ride upon. Not even the satisfaction or experience of earning it all.

That is, until his account was banned for good. Then, he had nothing.

 

[Vale]

Man, where is the fun in making a post using A.I.? I mean people can do whatever they want to do, and of course be handed the consequences as indicated here, but I'll never get it. I love writing up posts for partners, eagerly responding to what they write and have them in turn enjoy what I write. This is just crazy.

 

[Seravian]

This is absolutely fantastic to hear. People who use AI to write their posts have no place in a community full of talented people who pour their heart and soul into their writing.