We did it just to bug you. 
1 Like
I’m not criticizing the admin’s decisions because I do think they’re in a really tough spot (though I will say that I do really like @astrolub’s suggestion of having checkboxes on the submission form though since ideally we want it to feel easy for people to be transparent about their tool usage) but I do find the discussion that’s been happening, in this post and others in the past, very interesting because it feels like we’re orbiting the thorny philosophical question about whether the core purpose of the site is to be:
- A place for writers to share their work
- A place for people to get off
These mission statements used to be one and the same but now that machine can create passable prose and we don’t (ugh) “““need””” talented and enthusiastic writers around in order to have posts on the front page, it seems like they’ve diverged somewhat.
Users who think of it as a writing site do not want their work to have to share the site/compete for space with stories that were not written by a human. It’s low-key insulting, not to mention unfair due to the difference in the amount of effort involved/the pace at which someone can use AI to produce text.
Users who are more invested in the porn aspect of it have no problem with AI stories being on the site because content is content and they do not care about where it came from as long as it pushes their buttons. They find the apparent freedom that AI allows exciting and want to share the things they have it produce for them with other people who might be interested.
It’s two competing philosophies battling for the same space, which they have to do because GKS is the best space for text based kink content so there’s really nowhere else to go. In a perfect world people would keep their gen AI shit to themselves there would be a separate site specifically for people to post AI generated text so they don’t NEED to put it on the site with the stories people actually wrote, but this is unrealistic and problematic for so many reasons. I just hope that the (admittedly inevitable) compromises to the writing aspect of the site doesn’t drive away/demoralize writers.
5 Likes
The new rules are a step forward in keeping the site for writers, banning outright the AI stories that were produced quickly and with low to little effort, basically the kind of disposable stories that are generated on GenAI sites for a quick jerk. They have no place here, and I think we’d be better served by having a forum post or something that would show people how to make their own instead of reading someone else’s “generated smut”.
The only stories containing some AI we’ll accept here will be those where the author put in work and effort. I understand a lot of people think only the first type of AI stories exist, but that’s not the case.
Also, we rarely get stories of the first type submitted here now. And we probably won’t in the future since we are banning them. There’s not that many AI stories submitted now, even of the second type. Those stories are those submitted by some of the authors who commented on this very thread, and we should try to be respectful and not try to paint their work as low effort and non-human when you don’t know. And I don’t think it is their responsibility to justify their process to others.
In any case, don’t worry, trust in Corin. Can you imagine how much work it took to build this site and maintain it, all for free? Do you think he’d really let the site become just a repository for “generated AI smut”? He’s the one with the most to lose.
No. The tolerance shown seen here for some AI work comes from a motivation to help people tell their stories, and when used well, and again I know a lot of people are ignorant about this, AI can be a tool that boosts voices instead of replacing them.
You mention GKS as the best place to publish kinky stories. It became that in great part because of good decisions by Corin. So let’s continue to trust in those decisions.
2 Likes
As I’ve said, I have no issues with Corin or the decisions he made! I’m just pointing out some things I see interesting about the situation.
2 Likes
I understand the anxiety. I’m a software developer. Guess what task AI is getting better and better at? It’s still crap at it, but employers won’t see the difference before I’m fired and they face a catastrophe. 
1 Like
So after reading a lot of admin relies i think I’m even more concerned with the vagueness of the rules than I was before. It seems like the rules say and sound like the rules want to avoid wading into the ai discussion, but most of the replies are pretty heavily ai is the inevitable next step. That feels like a pretty heavy hand to one side. .
If I create a story/series, I create my characters always at first. A character cheat like used in pen&paper role play. All my stuff is “handmade” and only influenced by my dirty fantasy.
For me it is like growing a child. It’s a process of creating something. And this process needs time.
Often my writing is placed as a draft for weeks, or month. After a while I come back, review the written stuff, correct or change it and continue. All mistakes, (un)logical twists are mine.
I would’nt have this emotion if I had used an engine. (Thats my personal opinion)
3 Likes
To jump in with my two cents, I think that the new rule makes sense with the aim to dissuade the low effort AI generated stories.
AI is a tool and it can be used in good ways or in bad ways. I feel @Mafisto’s pain with regard to work and AI - as I am dealing with similar work issues in my job. However as a colleague pointed out, “AI is only so good, just look at the AI algorithm on Netflix and Hulu for suggested videos.” One small thing that I definitely hate about AI is how the em dash has become an indicator of AI prose. I am someone who has used the em dash for over 30 years in my writing of prose and papers, and to now see this glorious punctuation mark abused, it really, really hurts.
Despite all of that, AI is just a tool and I believe there are some valid uses for it in one’s writing, for example grammar, spell check, translations, and even to make sure that you didn’t name switch characters! I think maybe a check box for AI translated works is helpful as that is indicating that the work itself is written by a person but just in a different language and is translated - as this is slightly different than the AI Prose issue.
As for the AI Prose tag - it is tough to know where to draw the line and ultimately this is the difficulty. It is one of those things where you know it when you see it, but extremely difficult to define. Relying on the honor system is inviting some users to try and skirt it just to see if they can. You also don’t want to dissuade authors who use AI honestly, especially if they are just starting out in the craft.
So this leaves us with trying to find a happy medium like the AI Prose tag is trying to do. I am sure that this is a situation and policy that will continue to evolve with time. However as a stop gap and trying to meet the needs of the community of readers and community of authors I think it does hit an imperfect balance. At this point the approvers have given me no cause for me to distrust them and they are sincere in their efforts.
The one thing I would commend about the policy is that their will be communication with the author prior to the story being published and a story receiving the mark. My hope is that with honest discourse this will actually help the tag have a consistent meaning across stories, authors, and the site.
3 Likes
What I don’t understand is why, under the new rules, the AI-prose tag is optional.
I get that judging what is AI can be hard, but it’s one of the best features this site offered - protection from AI slop, and clear guidance on what was AI generated and what was human written.
That filtering was incredibly helpful. I have the AI tags blacklisted because I want to read stuff written by people. If you make the tags optional, you make it impossible to avoid AI work. Even if we put aside the ethical reasons why someone may not want to engage with AI work (the energy, the training, etc), I want to support authors putting in the work to write stories themselves.
This also does a disservice to authors. You’re effectively allowing things with AI sections and things hand written to be placed on the same site, with the same tags.
That seems deeply unfair both to readers and authors.
4 Likes
The short answer: we cannot be, and don’t want to be, the AI Police.
First, stories are free, approvers work for free. So there are limits to what you can feel entitled to. Everything has a cost and the load on free labor is too much to play police.
Second, we cannot create rules we cannot enforce fairly. And there is no way to know with perfect certainty if an author used AI or not. So when you cannot prove that a rule has been broken, you cannot enforce it. Even if a story is 99% likely to have AI in it, it would mean falsely tagging 1 story in 100. Most stories we detect as AI are more in the 50-70% likely to be AI, so do the math.
Not only that, but the more things you criminalize, the more criminals you create. The sort of atmosphere that was created when we attempted to police this was rather toxic. We depend on authors, they’re the ones who provide the content. Making them feel like criminals just to add some convenience for readers, especially when we could not truly with certainty prove anything, it’s just too much. So we’ll be doing our best to convince them to provide what readers would like, but no enforcement. Sorry.
Here’s the relevant part of the rules:
We don’t want to involve the site in anti-AI activism or spend resources to facilitate such things, but we do want to encourage authors not to frustrate readers. However, we don’t want to police terms that should remain between an author and a reader. Any sense of entitlement by readers to AI-use disclosure by authors makes no sense when the stories are offered for free. It is up to the author to choose to disclose or not. Still, we will strongly encourage authors to tag.
You can still filter the tag. We just cannot afford to give you a perfect guarantee and a flawless convenience.
And don’t forget that we will ban pure AI slop.
4 Likes
Fun fact. I have a book in publishing and my line editor added a ton of Em dashes. I use them a lot more now mostly because of the editor. I was not trained and never had a writing class so my methods are I think hair pulling for my editor haha. But hopefully using an em mark wouldn’t say oh look. AI. I was just someone who used “…” a lot and that’s how they replaced most of not all of it.
2 Likes
I appreciate the equitable approach, I guess. I will say I’m 100% prejudiced against GenAI, LLMs etc. and do not think there’s a way to ethically use the technology right now nor do I think anything produced on it is worth the time to read if the author couldn’t have been bothered to write it themselves.
If GenAI/LLM content isn’t wholesale banned (outside of obvious slop), being able to filter it out is the next best thing.
1 Like
You do you buddy , funny enough I’ve seen stories on the site that are covered in badges that indicate they are some of the most highly rated stories on the site and have tonnes of engagement in the comments . Guess what … They had a AI assisted tag . 
This discussion still seems to focus on people claiming not to use any AI vs those using some. And, the former not wanting the latter to be here, and the former claiming the latter put no effort or very little into what they produce.
So, just for doing so given the discussion…. I ran some of MY writing from even before I had any AI LLM/GPT/whatever to mess with… 4–FOUR!–of 6 stories that I wrote, one of which I wrote over the span of a few months…. were supposedly detected as being AI and with what I thought was a high percentage (70+). That was eye-opening for this discussion for me. I didn’t even know what these were or how to use those apps yet.
By contrast, I have been playing with a story idea and messing with a few different AIs with it. I spent over 3 hours yesterday just trying to get a few paragraphs written because of all the censoring, and refusal, and reinterpreting and redirecting. It really is NOT that easy. I still value the ideas it may provide, or its (partial) ability to analyze my logic. I would like to say my continuity; but, trying to have them actually stay with a story without distorting the history it has full access to…again, try it and you might understand if you are one that thinks it is all chocolate and bubble bath ease.
I don’t know what it is that would make me inhuman to the detectors… I have had people in my life who have questioned if I was human; but, I think that is different 
2 Likes
My tuppence as a very minor contributor (but I have been writing a lot of stories elsehwere for the past fortnight):
I enjoy playing with LLMs for fun, for roleplay, as a tool to quickly analyse some data and save some time.
But I cannot ever support LLMs supplanting the creative process of writing in a place where we share them.
Now, I’m hardly going to rage against whatever decisions the mods make and I’m hardly a major contributor of stories so even if I left tomorrow it wouldn’t make a lick of difference. But regardless… I prefer the AI-Assisted + AI-Created tags both existing.
the only real problem with them is a lack of disclosure on the author’s part as to what level of “assistance” is rendered. There’s a world of difference between “Hey Bot, take this page I’ve written and highlight some errors I made that a spellchecker wouldn’t” vs “Bot, write me some sentences for this story” or “bot, write this chapter for me”.
But, either way, the easy solution for me - and for most, Ithink - is to just block those tags entirely.
That said, I 100% agree with anyone who shows displeasure at the AI-made “stories” drowning out ones that take time to make. Like I said, I’ve been on a kick of writing stories and with my writing style of averaging 3.8-5k words per story, and that takes 4-6 hours depending on my level of inspiration, I would be pretty pissed to do all that work only to see 20 stories with those tags, knowing people put significantly less effort into them.
And of course I do fully understand that you can’t properly check if someone’s stories are AI-made, since AI “checkers” are notoriously unreliable, so an honours-system is the only way to enforce any rule.
Sorry if this was a bit rambly. TLDR: I personally don’t think any LLM involvement should be permitted, but I also recognise the impossibility of enforcing it outside of the honour system so I’m just gonna block the relevant tags.
- AI created stories have been banned . All that they have done is give a new name to AI assisted stories .
- If you count the number of AI assisted stories currently vs the non AI assisted stories I would say at least 5/10 - 1in favour of non AI assist
on any given update. The changes to the rules in March largely took care of that
3 What’s the point of trying to justify how much AI assist is in a story when those who wanna block AI assisted will do lt anyway .
1 Like
Out of curiosity, I put some of my text through an AI detector (I write everything myself, not using any AI other than spell check and occasionally I take Word’s advice on grammar). 96–100% human.
First, I need to tell some friends of mine who have their doubts as to my humanity. But that aside, as Mafisto pointed out, that could be interpreted to mean that 4 out of 100 stories would be flagged as AI when they’re human. I don’t think that’s quite the right way to read the stats, but more to the point, as @midwest_hypnosis pointed out, four of their stories were flagged as more likely than not at least partially AI-generated, when they weren’t.
So, as I noted in my reply a day ago, I certainly understand and appreciate the task the mods/approvers have to do and that they don’t want to be the police for this nor can they be when these detection tools are imperfect.
That all said, other than the time it would take to do the coding (which I have no idea about since I last did website development in the mid-2000s), I still think a checkbox as to how the AI was used that authors could fill out and then readers could filter on would be useful. E.g., I suspect many would filter out stories where large portions were written by AI, even if the author then went back and thoroughly edited them. However, I suspect many of those same users would not filter out stories where someone used AI to translate from another language or with some of the ideas but the text was wholly the human author’s work. The existing tags don’t capture that subtlety and I think it would be useful and used. It would be up to the authors to check those boxes, so the mods wouldn’t have to take extra time to do any follow-ups. Though, again, I don’t know how much effort it takes to code that.
Separately, I will also say that I empathize with jobs going to generative AI when generative AI stuff is still a fair bit of crap. My specific field in science has seen a massive surge of generative AI papers come out lately and most are utter crap. I know because I’ve reviewed about half of them because they tend to use my work as ground-truth. It’s very frustrating when people would come up to me at conferences a decade ago and say that because they disagree with one entry out of 500,000, they don’t trust my work and won’t use it, but they seem more than happy to use generative AI’s work that is perhaps 60–70% accurate and that’s only if you apply very loose matching criteria to the ground truth.
Check the reply’s buddy, they don’t care how AI is used , they don’t believe can be used in a positive way and they will block it regardless . They don’t want to know . you can’t please all the people of the time .
1 Like
I’ve read all the replies. I am making a new suggestion that they might consider for the future, given that the AI policy has now radically changed twice in the last month. I am also saying that this could be a user-facing feature that could be useful and would not necessarily need to be policed. As @d-boi and others have said (including me!) there’s a huge difference between AI = spell-check versus AI = rephrase sentences versus AI = translate versus AI = write me a few chapters on this topic. The current tags don’t capture that, while I think many readers and writers would appreciate a method to capture that. If those who know the backend are able to implement it – free/volunteer time- and technology-wise.
1 Like
Most of the people who dislike AI will not care about those differences and will block them anyway . And actually I don’t believe there have been hardly any stories in the last few months where anyone has written a few sentences to generate whole stories or chapters I honestly don’t believe that’s how most AI stories are made now
1 Like