The rules for AI involvement in stories

142

Stories that were banned here for what reason? Because they violate our rules, which are certainly rather specific (especially the no-realistic-rape rule)? I can imagine that these would get a good reception elsewhere.

That we ban these has nothing to do with their quality or reception. It’s a problem with my own conscience. I’m going to admit that I’m a bit of a hypocrite there myself. But there are so many bad things happening in the world, where people are raped and killed, I don’t need to have this glorified in stories on my site.

But if you’re talking about stories using AI: We haven’t banned a single story yet because of AI usage. In fact, the first story that would fall into the category of not being published due to noticeable and overly distracting AI prose is just in the approvers’ queue, and we’re about to tell the author why we can’t publish his story.

143

Ok, we’ve been talking with, and sometimes past, each other about theories and potential future scenarios, and our personal beliefs in a lot of stuff, and it’s been interesting as well as frustrating for everyone due to sharp differences in opinions.

I want to just bring to your attention how things actually changed in practice. @PupVince that might answer your question in the other thread.

Trained people evaluate each story submitted to see the likeliness of AI, exactly like before.

When a story is what I call a “genie-wish” story, the result of describing your kink to AI, like “tell me a story about a man visiting a house and hypnotizing the real estate agent into having sex with him in the master bedroom” and poof, getting a full fledged story, it is quite easy to tell and the resulting story is usually awful enough that even if a human wrote it, we easily could decide not to publish it anyway. Instead of sticking an “ai created” tag on this story, we simply ban it. These stories have strong AI patterns all over the text, they really stand out to a trained eye.

When a story shows a high degree of AI patterns in it, but has been reshaped through the creative vision of a human, not just wish fulfilling, there is a high likeliness, though never a certainty, that the story has been assisted through AI. We write to the author in the same way that we did before, telling them that they should tag it, exactly like before, the only difference being that we do it in an annoyingly insistent way instead of just outright enforce it without proof. In the rare cases where the author insists that there is no AI and doesn’t want to tag it, we tell them “I wouldn’t do that if I were you”, explain all the reasons why, but in the end, we’ll accept their decision because we have no way of being sure. This will happen quite rarely because it is not actually in their interest to do that. Authors want good feedback and hate bad feedback. What we detect in the text, readers who hate AI patterns will too, and although we wouldn’t let readers make unfounded accusations, even a single mild comment of “I thought the writing in the following paragraph had a lot of similarities to bad AI writing.” can motivate an author to tag so the readers likely to post this bad feedback will instead filter their stories out.

Right now, we have an example of both kinds of stories in the queue. So business as usual.

2 Likes
Setting an example for authors to disclose their AI use
144

Must admit I would love to hear people’s views on this ( unless of course there is already a rule that should mean this wouldn’t happen ) in which case ignore me

the reason I link this to the AI debate is because if someone says they couldn’t forgive reading a story that’s had AI involvement and was untagged even if they enjoyed the story, would they be as unforgiving if it was human written ? ( also to be clear not talking plagiarism, I’m talking ghost writers / commissioned pieces where the writer has closely worked with the person who has been specific in what they want them to write )

remember I am in favour of AI tagging because it’s right to be transparent as possible. I’ve already expressed clearly why I decided to publish stories rather then keep them to myself .

145

I would absolutely feel cheated and upset by that if it were done without credit, but thankfully I don’t know of any cases like that (here in the gay MC erotica world, at least – famous real-world cases like K.A. Applegate have stuck in my craw since I was a child).

I do know of cases here where it happens with credit, like with MonsterMash62/Hugh Michelsen’s Married Life that was ghostwritten by ChaoticDjinn, and while that is substantially better, I do still feel slightly annoyed at the correct credit being buried in the text and a false one being how stories like this are actually sorted and findable. (In real life, see also: almost every celebrity autobiography, where I have the same feeling.)

As an irrelevant postscript, I want to also clarify that these feelings don’t apply to cases where the original author was originally credited but then closed their account and permitted the stories to stay on another person’s account, like Swizzington’s that are sitting on Ediam’s. That’s a situation that’s sad regarding the fact that the person left, but totally fine in terms of crediting since the software allows the story titles and author profile to make it very clear what’s happening.

2 Likes
146

Thank you for replying buddy , I respect you and your point of view on these subjects, I would never stop tagging my stories as a AI assisted author out respect for the community, hopefully things settle down and all authors and readers can co exist and get what they want from being part of this wonderful community.

147

It seemed that was the point being made… that there are more AI happening than not. And the Human-Only are being drowned out. That’s why I thought if an overall tag like exists at the top for hypno only or not, for example, existed…they would ONLY see the Human-Only. And, there would be less need for detection and determination how much was AI or not and having those people have to defend themselves and how they used AI as well. But thank you for acknowledging and sharing. As someone else said, we don’t know all the closed-door discussion already having occurred. I was also pointing more at the people wanting to stand out more as having not seemed interested in their own work carrying a label or identifying mark.

1 Like
148

I do still think, genie being out of the bottle and all, that some documentation of how AI is involved in stories would nice. Not for the purposes of avoidance but for the purposes of historical record and discussion. I admit Tags might not be the way to go about it though. /pos

2 Likes
149

Hehe if you want historical record. I wrote a whole story about the process I went through and exactly how AI was involved and how I used it in each of my stories to date . It was actually rooted largely in truth if you ignore the evil AI voice in my head :face_with_spiral_eyes::face_with_spiral_eyes::face_with_spiral_eyes::face_with_spiral_eyes::face_with_spiral_eyes:

1 Like
150

As opposed to your analogy, which was so much better because it exaggerated the people you don’t like to distort them and further dismiss what they’re saying? Lol

I think just think a lot of people are being very uncharitable to the anti-AI crowd. I can’t speak for everyone, but personally, it’s not that I don’t understand the points the pro-AI crowd are making, it’s that I don’t agree with them. I have read every comment in this thread and I get what people are saying about there being many use cases for AI and that there are people who find it helpful, but I have not seen anything that actually convinces me that the positives of (specifically generative) AI outweigh the negatives.

Also the reason I’m ignoring the idea of a “Human Only” tag is because I believe it to be unnecessary. It would be like tagging a story “contains words…” of course it does! It’s a story, what else would it have? The people using AI are the people deviating from the standard (after all, the vast majority of works on the site are already “Human Only” and it seems like submissions are still the same) so they’re the ones who should have to tag.

2 Likes
151

Not to be pedantic but nobody asks human writing authors to tag that a story has been written wholly and exclusively by them, it’s assumed and why wouldn’t it be . But it’s not an impossibility that an author could publish a story written for them by another human using the detailed directions they have given them and like the guy above said , some people would feel very annoyed to be deceived in that fashion . But it doesn’t discount the slim possibility though .

152

The funny thing is that most of the “pro-AI” crowd (I guess you count me among them, don’t you?) are very much aware of the dangers and possible destructive force AI might have on society. I think I’ve written that multiple times.

If I could, I would simply snap my fingers to undo all AI technology.

However, I don’t have that power. Just like the people in the early 20th century, who were opposed to those stinky, fuel-driven automobiles (and rightly so, if you look at the outcome), couldn’t stop them from taking over the world.

We are not going to stop AI. It might fail on its own accord, but I think hoping that is unrealistic. You can decide not to use it, and that’s totally ok - you can still use your old horse carriage for as long as you can. But you cannot keep other people from using the same streets with their dreaded, stinky, loud cars, where you just want to ride your carriage in peace.

So we have to learn to live with the technology and adapt to it as much as we can. There is simply no other choice.

You can try starting a new GKS-like server and ensuring there is no AI content on this server whatsoever. I wish you good luck with that, but I’m pretty sure you’re going to fail because you will end up in countless fights with people who lie about AI usage.

To make it clear, before people misunderstand this again: Anyone is welcome on this server, even if they despise AI and want to block it. But they have to respect that I made a decision for how we deal with AI content, which, in my opinion, is the best compromise I can come up with.

4 Likes
153

AI is like…

The Pandora Box
or
A Mogwai (aka Gremlin) — once in touch with water or feed after midnight, the moster will be free.

(Sarcasm)

And about the fuel driven vehicles… they will vanish

1 Like
154

Yeah all those non believers will regret it when Ultron takes over :joy:

1 Like
155

As you saw in my other post, I did both in my examples, a tag AND a rather long disclaimer in the story itself explaining how AI was used for the story. We will encourage authors to do the same. There’s a lot of ignorance and preconceptions about AI, so it can help at least inform people. Probably won’t change many minds, but at least they’ll be more informed when they stick to their guns.

I don’t consider myself pro-AI. I never was pro-social media. When I got my iPhone in 2006, I would always forget to bring it with me when I went out. I’m still rarely on my phone, it’s like I get asked for an Auth number and I groan because my phone is in the other room. I saw most of the problems social media would bring, never understood why schools ever allowed students to have their phones in class. Still, I enjoy the site we have here and my FoundryVTT to play Dungeons & Dragons online with my friends all over the world.

I may be less anxious about AI because as much as I hear the hype from companies, I know enough about it to clearly know its limitations: it won’t be able to do 80% of what it’s being sold as being able to do, there’s basic limitations inherent in it that it will never get over, unless another major breakthrough arrives in a decade. Don’t forget that AI stalled for decades before this sudden revelation we got about LLMs surprisingly doing much more than we’d ever expect them to do. So I’m not against AI or pro AI, I’m used to be reasonable about any technology I use, there are some interesting things it can do that I’ll use it for but I will help as much as I can to not let it invade spaces with its worst aspects.

3 Likes
156

yeah that’s mostly what I was thinking of! One can only hope other writers will be as open about their processes with this new tech.

4 Likes
157

So quick question for the group: Would you consider any of the following uses to be significant use of AI?

  • Using it for spell check and/or grammar check
  • Using it to check story consistency - for example ensuring that there are no name swaps, change of spelling of names, or incorrect locations
  • Using it to check for flow - making sure that it is internally consistent
  • Assistance in rephrasing of a sentence or very small passage that is awkwardly phrased

In your opinion would a story that used AI for these purposes require the AI tag? AI can be helpful as an editing tool. @Mostly_Ghostly - I believe that these would be positive uses of AI tools (with maybe the exception of the last bullet point, which even I would say is debatable). Is it perfect? No, but it can be helpful to get past that initial draft phase. Is it helpful for writers who don’t have anyone to edit their stories? Yes. In these cases, I don’t believe that AI tag would be appropriate.

On the other hand, if the author is using the AI to write the bulk of the story (or significant portions) then changing what the AI produced, that is a different matter altogether and something I would say the AI tag is appropriate. Or using a variety of prompts to have the AI craft a story - that is

In general - I am AI neutral. I think it is a tool and like any tool it can be misused. Examples of misuse: some universities are using it to help categorize and summarize student applicants, employers are using it to screen job applicants, tools that use it to help users select their medical insurance - these are uses which I think are going to ultimately cause more harm than good.

In my job I am actually involved in developing AI tools and chatbots and much to the chagrin of some of my colleagues I take a very hard line on setting limits on what the tool can and should do. In testing I judge its out-put very harshly. Even when speaking with potential clients, if I am demoing the tools I always put it in perspective of the fact it is only a tool and does not replace the need for users to review and understand specific materials or replace their need to make decisions and think. As @PupVince pointed out, “just like mogwai, if you keep feeding AI, after a certain point it turns bad.” I think that this is applicable for both quality and morally.

Finally - @Mafisto - totally off topic - there seems to be a lot of people on the site who play D&D, why haven’t we organized D&D game nights yet? And also Foundry VTT rocks and if you haven’t used it yet (and if you are DM) I highly recommend Dungeon Alchemist for map making. Don’t worry it does not use AIs. :slight_smile:

158

The answer to your question about the AI tag is in the rules, of all places? :slight_smile:

Tagging

The only type of AI use that will be tagged is the only one that can be detected easily by readers and approvers. It’s the one that bothers readers the most, so it’s the most useful one to filter. We will use the “contains ai patterns” tag to indicate that the story contains passages that read like AI generated text with its signature features.

So the tag is not based on how you created the story or how you used AI or anything, it’s based on the story itself, if it contains clear AI patterns in it that readers don’t like as a style and want to filter out.

We also encourage authors to disclose other uses of AI and be as detailed as they like in a disclaimer directly in the story, as I did with the ones I offered as examples. Corin may add something to facilitate adding any type of disclaimer, not just AI use ones, if he ever gets to it.

So none of the scenarios you mention could potentially result in significant AI patterns and thus the tag. There are other methods that may though if you overuse them, but if it’s only here and there and you edit what the AI gives you, it shouldn’t show in the story itself. But when most of the story was generated, usually those patterns will show and be pretty hard to get rid of, so the tag would fit for these stories. And of course, if it’s a story completely generated from AI with no human creative vision or editing, we will just ban it, won’t be published. People who translate using AI often get enough of those patterns to be encouraged to tag, it’s better not to use AI for translation.

159

As for D&D, I’ve started playing since 1980, I’m a forever DM. I have three games a week, all different campaigns I created, with people from Quebec, France and Switzerland on video webcams. We use FoundryVTT and yes, I use Dungeon Alchemist and Dungeon Draft for maps. We record our games privately on video on YouTube and since 2014, there are more than a thousand sessions there. All in French of course, my first language.

2 Likes
160

I think his question was less targeted at us, but more at the group of people who wants to block any kind of AI usage in stories. At least that’s what I got from @Ayebarra’s message. Am I right?

161

Sorry, I may have misunderstood that it was a purely philosophical question, not a rule clarification.