AI or LLM

Status
Not open for further replies.
Engagement is low here - if you mean comments and votes. But views elsewhere are a joke. You might as well set up your own blog.

Nope - viable as in having people actually read your story.
I'm not going to pretend I know anything about the process, but it seems like an archaic solution to a modern problem. If you're going to maintain such a hardline stance against AI, then just auto-reject every story that scores above a certain probability of AI-usage. I don't think talking about this is productive though. The impression I've gotten is that this site hasn't adapted particularly well over the years.

I also have no idea what can be considered a viable reader-count as I'm fairly new, and only write niche content. However, the story I tried to upload here (probably around a week ago, got rejected yesterday), has already received about 20,000 views altogether on the other sites I uploaded to. Take that with a grain of salt, as I know bots are pretty prevalent everywhere.
 
Good on you for interceding. I can definitely respect that. However, most new writers simply don't have the time or energy to waste on trying to convince an AI or mod that they themselves didn't use AI. That time and energy is better spent elsewhere.
If they're writing erotica, this site is far and away the biggest publisher - no other site comes close in terms of readers.
I'm aware my claim to leave is irrelevant, but that wasn't the point I was trying to make. The point is that nearly every new writer here will be in the same situation, and face the same decision. That's not insignificant.
200 or so stories get published every week. There is no lack of content. Of those, probably 70% - 80% are likely to be one-off stories, or the first chapter of a limited number of stories, with a smaller group of writers submitting repeat content.
At the very least, the owners of the site could consider running every submission through their AI checker automatically, and rejecting them automatically if AI use is suspected. At least that way, writers don't have to wait 4+ days.
That's the process - a bunch of bots look for various things, resulting in rejections in a fairly quick time (hours). Then the queue goes onto a conveyor belt that gets in front of human eyes (typically, 2 - 3 days).
As for other viable places to publish... I've already found five. Of course, the content I write is niche fetish content, so there's that. You may not be looking hard enough.
What's the readership volume of those sites compared to Lit?
 
...a ton of new, talented writers...

You keep saying that. Where are these hoards of talented, capable, and I presume young, authors who are writing erotica? I'm waiting. If they were talented, as you say, they'd actually be good at writing and wouldn't be leaning on synthetic prose generators.

...niche content... received about 20,000 views...

Doesn't pass the smell test. "Niche" is usually that way because it isn't common or popular. So name the sites.

I know bots are pretty prevalent everywhere.

You said it, not I. On the Internet, nobody knows you're a dog... or a bot.
 
I completely agree with this sentiment. Websites like this are extremely useful for new writers. I just worry that a lot of new writers are getting their works flagged as AI when no AI was used. It's already pretty well-known that AI detection tools are extremely flawed. I strongly believe that this site is putting itself at a huge disadvantage in that regard.

I worry about this too, and I'd like to see more transparency about the flagging process - I can understand why Lit might not want to tell people what software they're using, but even just some numbers on what % of submissions are flagged (and what percentage of those are overturned on review?) would make this a more informed discussion.

But I'm not sure there are any good solutions to the current situation. Auto-detecting AI content is unreliable, manually flagging it is a ton of work and still unreliable, letting it all through is bad too for reasons I mentioned above, and drone strikes on LLM data centres are illegal.
 
If they're writing erotica, this site is far and away the biggest publisher - no other site comes close in terms of readers.

200 or so stories get published every week. There is no lack of content. Of those, probably 70% - 80% are likely to be one-off stories, or the first chapter of a limited number of stories, with a smaller group of writers submitting repeat content.

That's the process - a bunch of bots look for various things, resulting in rejections in a fairly quick time (hours). Then the queue goes onto a conveyor belt that gets in front of human eyes (typically, 2 - 3 days).

What's the readership volume of those sites compared to Lit?
I have no idea what amount of readers is acceptable for vanilla erotic content. I don't read or write vanilla. Honestly, I'm probably not approaching this from an ideal perspective because of that.

A huge amount of readers of erotic fiction are looking for something very specific. It's not the same as an epic fantasy fan reading epic fantasy. They need certain erotic beats hit and certain kinks satisfied. So, while 200 stories a week is a decent number, a very small fraction of those appeal to individual readers.

The story I wrote that got rejected for suspected AI use was pending for almost a week. If a bot flagged it for AI use initially, having it rejected immediately would have been preferred. At least then I may have considered contacting a mod or taking another look at it. I've since moved on to other works though, and really couldn't be bothered.

Of the sites I use, and in the week since I published my work, two of them have generated reads in the mid-100s, two in the low-mid 1000s, and then Deviantart (which I know is a shit-show, but non-vanilla/fetish content does really well there).
 
You keep saying that. Where are these hoards of talented, capable, and I presume young, authors who are writing erotica? I'm waiting. If they were talented, as you say, they'd actually be good at writing and wouldn't be leaning on synthetic prose generators.



Doesn't pass the smell test. "Niche" is usually that way because it isn't common or popular. So name the sites.



You said it, not I. On the Internet, nobody knows you're a dog... or a bot.
That's the thing though. These AI detection tools are extremely flawed. A lot of new writers waiting days for a check here only to receive a false flag will simply go elsewhere. They aren't using AI, and the frustration they feel from that days-late false flag will drive them away.

Believe it or not, niche EROTIC content often does quite well (sometimes better than vanilla), as readers often seek very specific plot beats and kinks. I write size-play/size-change, power dynamic, role-reversal transformation stories that often include futa. It's incredibly niche, but there are many readers and very few quality writers.
 
If they're writing erotica, this site is far and away the biggest publisher - no other site comes close in terms of readers.

Not sure about that. It's hard to find a good source of stats for comparison, but Literotica has had approximately 64k new stories in the last year and hosts about 750k total.

Meanwhile, for 2025 Ao3 reported approximately 100 million page views a day and over 400k comments a month: https://archiveofourown.org/admin_posts/33966

I don't know exactly what the average view rate is for Lit stories, but it wouldn't come anywhere close to 100 million views a day. And while Ao3 isn't all erotica, the percentage is high enough that I suspect they end up being a bigger erotica publisher than Lit.

This is not to say that Ao3 renders Lit obsolete; they're different sites serving largely different audiences. But Lit isn't the only big fish in town. One thing Ao3 has going for it is that the design is much more scalable than Lit's, but then I don't know whether L & M ever aspired to infinite growth. It may be they're perfectly happy not having the site grow faster than it is.
 
I worry about this too, and I'd like to see more transparency about the flagging process - I can understand why Lit might not want to tell people what software they're using, but even just some numbers on what % of submissions are flagged (and what percentage of those are overturned on review?) would make this a more informed discussion.

But I'm not sure there are any good solutions to the current situation. Auto-detecting AI content is unreliable, manually flagging it is a ton of work and still unreliable, letting it all through is bad too for reasons I mentioned above, and drone strikes on LLM data centres are illegal.
Ya, taking such a hardline stance on AI use puts Lit in a pretty tough position, especially since they don't seem to want to be transparent about their process.

I'm only writing all these posts to vent a little, as I waited for a week before receiving a rejection for potential AI use. I didn't use AI, and that response left me feeling frustrated and bitter.

I've uploaded to other sites, and have received quite a lot of reads and engagement. It puts me in a position where I may just not bother with Lit in the future. I imagine a lot of other new writers feel similarly.
 
Ya, taking such a hardline stance on AI use puts Lit in a pretty tough position, especially since they don't seem to want to be transparent about their process.

I'm only writing all these posts to vent a little, as I waited for a week before receiving a rejection for potential AI use. I didn't use AI, and that response left me feeling frustrated and bitter.

I've uploaded to other sites, and have received quite a lot of reads and engagement. It puts me in a position where I may just not bother with Lit in the future. I imagine a lot of other new writers feel similarly.
What tough position? They host stories here without making money on those stories. It's not like they're losing revenue if a subset of people don't publish here due to frustrations with the submission process. This website, insofar as I can tell, is a passion project for a couple people who appreciate human-made erotica.

It's totally up to you if you want to publish here or not. I get that the rejections are frustrating, or feel personal. I've had it happen to me a couple times. But the thing is, they aren't personal at all. In fact, they're about as impersonal as you can get. If you've found other sites, that's great. It's a pros vs. cons thing. What are the pros of putting in a little extra effort to get your story published here vs. not bothering? If it's not worth the hassle to you, that's up to you, my friend.

Most of the people publishing here aren't doing so to make money. Their main motivation is to share their works with a community that will read and value it. In that light, Lit's instinct to want to protect that spirit makes perfet sense, because while it's disheartening to get a rejection for suspected AI use, it's a hell of a lot more disheartening if nobody ever reads your story because it's buried under a mountain of AI slop.
 
What tough position? They host stories here without making money on those stories. It's not like they're losing revenue if a subset of people don't publish here due to frustrations with the submission process. This website, insofar as I can tell, is a passion project for a couple people who appreciate human-made erotica.

It's totally up to you if you want to publish here or not. I get that the rejections are frustrating, or feel personal. I've had it happen to me a couple times. But the thing is, they aren't personal at all. In fact, they're about as impersonal as you can get. If you've found other sites, that's great. It's a pros vs. cons thing. What are the pros of putting in a little extra effort to get your story published here vs. not bothering? If it's not worth the hassle to you, that's up to you, my friend.

Most of the people publishing here aren't doing so to make money. Their main motivation is to share their works with a community that will read and value it. In that light, Lit's instinct to want to protect that spirit makes perfet sense, because while it's disheartening to get a rejection for suspected AI use, it's a hell of a lot more disheartening if nobody ever reads your story because it's buried under a mountain of AI slop.
Ya, I understand most of this isn't profit driven. However, there is still a market-share issue. As AI develops further, and other new or existing platforms adapt more quickly/efficiently/effectively, Lit will lose the market share they currently have. That means less writers for people to read as they move onto better platforms, and then less readers for the people that do publish here as those readers follow the writers.

I also understand that none of this is personal, and that isn't really what I'm upset about. I'm most upset about the fact that I was made to wait a week only to be notified my piece was rejected because of suspected AI use. I'm well aware it takes mere seconds to run a story through AI detection software. If it had simply been rejected within a few hours, I'd have still been in the headspace to edit it or try to appeal. Yet, I've finished another story since then, and have already started on another one after that. I can't imagine many new writers are going to be willing to sacrifice the time and effort they should be putting into improving their craft just to satisfy Literotica's AI filter.
 
I can't imagine many new writers are going to be willing to sacrifice the time and effort they should be putting into improving their craft just to satisfy Literotica's AI filter.
You're saying that as if working on the story to get it to pass the AI filter isn't a way of improving your craft in the first place. But that's a false dichotomy.

Whenever you rework a story, you're putting in time to improve your craft, because you're looking at the story again and analyzing what is and isn't working, what could be improved upon, what needs to be expanded, what needs to be cut. Editing and revisions are as important an aspect of the craft of writing as the initial writing process, and you can actually learn more from that than you can from simply pumping out rough drafts and doing minimal edits. Speaking from a lot of experience of pumping out rough drafts with minimal edits when I first started writing. I only really started getting better once I really dug in and tried to understand writing on a much deeper level, seeing what I was doing well, what wasn't hitting right, where I was missing opportunities. Trying out different styles and phrasing, structures, beats, all of it, as part of my drive to truly get what makes writing good.

When my story was flagged for AI, twice, I looked it over and realized it was too uniform in its structure and needed more variety. The story that came out from those edits was a big improvement over the initial submission.

I can't speak as to why it takes as long as it does to get that filter to kick your story back. The process is obtuse, which leads to frustration, which is understandable. But just because your story is flagged as AI doesn't mean that reworking it to get it past the filter is a waste of time. Reframe it, and think of it as an opportunity to improve your story.

It's up to you if you want to go back and edit something. Plenty of people work on multiple projects at the same time, or go back and edit things they "finished." If you don't want to, that's fine, it's a pure judgement call about where you want to focus your time and effort, because I get that we don't have enough time to do all the things we want to.
 
You're saying that as if working on the story to get it to pass the AI filter isn't a way of improving your craft in the first place. But that's a false dichotomy.

Whenever you rework a story, you're putting in time to improve your craft, because you're looking at the story again and analyzing what is and isn't working, what could be improved upon, what needs to be expanded, what needs to be cut. Editing and revisions are as important an aspect of the craft of writing as the initial writing process, and you can actually learn more from that than you can from simply pumping out rough drafts and doing minimal edits. Speaking from a lot of experience of pumping out rough drafts with minimal edits when I first started writing. I only really started getting better once I really dug in and tried to understand writing on a much deeper level, seeing what I was doing well, what wasn't hitting right, where I was missing opportunities. Trying out different styles and phrasing, structures, beats, all of it, as part of my drive to truly get what makes writing good.

When my story was flagged for AI, twice, I looked it over and realized it was too uniform in its structure and needed more variety. The story that came out from those edits was a big improvement over the initial submission.

I can't speak as to why it takes as long as it does to get that filter to kick your story back. The process is obtuse, which leads to frustration, which is understandable. But just because your story is flagged as AI doesn't mean that reworking it to get it past the filter is a waste of time. Reframe it, and think of it as an opportunity to improve your story.

It's up to you if you want to go back and edit something. Plenty of people work on multiple projects at the same time, or go back and edit things they "finished." If you don't want to, that's fine, it's a pure judgement call about where you want to focus your time and effort, because I get that we don't have enough time to do all the things we want to.
Oh, I definitely agree that editing is an essential part of the process, and should be practiced extensively. However, the idea of going back to a story because an AI filter came to the conclusion that it was AI, and then editing it to satisfy that filter feels pretty gross.

I'm going to assume you didn't use AI for your story. It was completely original, and yet the filter failed twice. Going over it again may have made you a better writer, but that isn't the issue here. The issue is that your work, which was entirely original, was rejected because their filter failed. Twice. That alone illustrates that the filter is extremely flawed.

Why would I want to go back and edit my work with the intent of getting it past a flawed filter? That's ridiculous.
 
I dare say that in time GenA writers will start to imitate AI. We already have assholes in business telling us that 'We need unilateral growth across all verticals in order to align on the key deliverables'. These are the same people who text from a poolside to say you're entire department has been replaced by Grok.

Personally I like the idea of unilateral growth on verticals - I'm always happy to align those deliverables into my box. :devil:
You make an interesting point. Do you think writers will eventually adapt their writing to ensure there's no way it can appear as an AI creation? Do you think AI will eventually adapt to that? Do you think the AI filters will then adapt to police that new adaptation?

Honestly, writers making changes to their work to satisfy and convince an AI that their work isn't AI feels pretty distasteful.
 
Oh, I definitely agree that editing is an essential part of the process, and should be practiced extensively. However, the idea of going back to a story because an AI filter came to the conclusion that it was AI, and then editing it to satisfy that filter feels pretty gross.

I'm going to assume you didn't use AI for your story. It was completely original, and yet the filter failed twice. Going over it again may have made you a better writer, but that isn't the issue here. The issue is that your work, which was entirely original, was rejected because their filter failed. Twice. That alone illustrates that the filter is extremely flawed.

Why would I want to go back and edit my work with the intent of getting it past a flawed filter? That's ridiculous.
I'm not saying to edit it to simply to satisfy the filter. If it gets flagged, it's an opportunity for you look at your work, determine if it can be improved and what those improvements are, if any, and make them or don't. If you think it's good as is, you're free to add a note to admin affirming you didn't use any AI in the crafting of the story. Somtimes that's sufficient to trigger a manual review, and it gets passed through. If it's still flagged as AI, then don't publish on it on Lit, see if you can get it somewhere else.

I'm in no way suggesting you compromise the story to try to meet some nebulous criteria that, frankly, none of us know what it is, so how can we even pander to that filter? But it's an opportunity to see if there are areas for improvement. Whether or not you decide to take it is up to you. If I thought my work was good as if, but still couldn't get past the filter, I wouldn't dream of modifying it to meet some hand-wavy criteria I know nothing about, because the story is more important than my ability to publish it on Lit. I could take my ball and go elsewhere, free internet. But if I look at it and go, "Huh, I guess maybe the writing is a bit stiff. I could vary the sentence structure a bit, lean on fewer cliches, make the dialogue more natural," what's the harm in polishing your work for the sake of making a better story? So what if the trigger was being rejected by AI, the goal of editing is to improve the story, not compromise it to fit into someone's neat hole.

It's a game of trade-offs. You have a more lenient filter, more AI slop gets through, fewer human writers get exposure. It would be nice if we lived in a perfect world and we didn't have to deal with this, but we don't, so it's a matter of what the site feels is the best approach to best fit their vision for the stories that get through.

We can argue about what the best approach is 'til we're all blue in the face and all I want to do is put up Numento's "Wrath" anytime somebody enters the AI-talk death spiral for 20 pages' worth of thread, but what good is that going to do us? The T&C are the T&C. Our opinions on the matter don't really amount to squat. The admin are free to do with the site as they will. Not saying it isn't frustrating, or not annoying, or even good policy, but it's the way things are, and until they decide to change things up, we're stuck with it. So writers either find ways to work with the system, or decide that the hassle of dealing with this system isn't worth it and take their work elsewhere. Totally up to each of us to make that determination and how much effort we're willing to put in.
 
I dare say that in time GenA writers will start to imitate AI.
It's already happening. I see it on websites that claim not to use AI using terms that are clearly AI-coded (sometimes it's clear they're lying about not using AI, other times it's clear that someone is just imitating AI, because the writing is too sloppy to actually be AI). Not written in the same way that screams AI, but borrowing AI-isms. "Why this matters." "The quiet thing they don't want you to know." It's very disheartening that everything is converging toward a single bland way of writing, where all our voices are melding into a single distorted belch of meaningless drivel and half the sentences are accentuators that don't actually say anything.

And it's not just genA, it's way more pervasive than that already. My friend and I send screenshots of news articles (mostly me) and videos and social media posts (mostly her) that have very obvious AI-coded terms and phrases, and it's now multiple times a day.
 
I'm not saying to edit it to simply to satisfy the filter. If it gets flagged, it's an opportunity for you look at your work, determine if it can be improved and what those improvements are, if any, and make them or don't. If you think it's good as is, you're free to add a note to admin affirming you didn't use any AI in the crafting of the story. Somtimes that's sufficient to trigger a manual review, and it gets passed through. If it's still flagged as AI, then don't publish on it on Lit, see if you can get it somewhere else.

I'm in no way suggesting you compromise the story to try to meet some nebulous criteria that, frankly, none of us know what it is, so how can we even pander to that filter? But it's an opportunity to see if there are areas for improvement. Whether or not you decide to take it is up to you. If I thought my work was good as if, but still couldn't get past the filter, I wouldn't dream of modifying it to meet some hand-wavy criteria I know nothing about, because the story is more important than my ability to publish it on Lit. I could take my ball and go elsewhere, free internet. But if I look at it and go, "Huh, I guess maybe the writing is a bit stiff. I could vary the sentence structure a bit, lean on fewer cliches, make the dialogue more natural," what's the harm in polishing your work for the sake of making a better story? So what if the trigger was being rejected by AI, the goal of editing is to improve the story, not compromise it to fit into someone's neat hole.

It's a game of trade-offs. You have a more lenient filter, more AI slop gets through, fewer human writers get exposure. It would be nice if we lived in a perfect world and we didn't have to deal with this, but we don't, so it's a matter of what the site feels is the best approach to best fit their vision for the stories that get through.

We can argue about what the best approach is 'til we're all blue in the face and all I want to do is put up Numento's "Wrath" anytime somebody enters the AI-talk death spiral for 20 pages' worth of thread, but what good is that going to do us? The T&C are the T&C. Our opinions on the matter don't really amount to squat. The admin are free to do with the site as they will. Not saying it isn't frustrating, or not annoying, or even good policy, but it's the way things are, and until they decide to change things up, we're stuck with it. So writers either find ways to work with the system, or decide that the hassle of dealing with this system isn't worth it and take their work elsewhere. Totally up to each of us to make that determination and how much effort we're willing to put in.
Any writer worth their salt is going to edit their story, regardless of whether an AI filter tells you to or not. I haven't published on Lit yet (first story got rejected), and probably won't given how strict their filter is, so I don't really have any skin in the game. I'm really just venting and making observations.

Imagine being a new writer. You've edited your story and uploaded it on a bunch of different platforms. You're getting decent engagement everywhere but on Lit, which hasn't even reviewed it yet. You continue writing, and release your next work, which also gets engagement on other platforms. So, you start on your next. Then you receive a message from Lit saying your first story was rejected. Most writers in that situation will just move on from Lit.

Now imagine an experienced writer who's built up a pretty dedicated following on multiple platforms, Lit included. Editing and publishing has become a smooth process, except on Lit, where you often have to make small tweaks to satisfy their AI filter. Most established writers will also eventually give up on Lit in this situation.

The issue is that this will consistently push writers (and thus readers) to other platforms. I don't know when it'll happen, but this will eventually lead to this websites death spiral.
 
Ive played with GenAI a lot, mainly for music (Suno) and "art" (Civitai, ComfyUI). By comparison with fiction-writing, the slop I can create there is pretty good.

But when it comes to creative writing, I'm surprised at how poor the quality is, when you consider that pure language models are the most well-developed and succesful of all of them.

As to this site, one solution would be to force an #AI tag on stories that the (often wrong) detector has flagged.
But to be honest, I suspect that the site owners are using AI detectors to triage the stories , mainly to reduce the human workload of actually reading too many submissions every day.
 
Ive played with GenAI a lot, mainly for music (Suno) and "art" (Civitai, ComfyUI). By comparison with fiction-writing, the slop I can create there is pretty good.

But when it comes to creative writing, I'm surprised at how poor the quality is, when you consider that pure language models are the most well-developed and succesful of all of them.

As to this site, one solution would be to force an #AI tag on stories that the (often wrong) detector has flagged.
But to be honest, I suspect that the site owners are using AI detectors to triage the stories , mainly to reduce the human workload of actually reading too many submissions every day.
There's just so much to keep track of, even with short stories. Narrative consistency, tone, style, character backstories, character voice, character arcs, consistent themes, worldbuilding elements etc. Holding it all together requires a lot of processing power. I'll admit I don't really know much about AI though, so I could definitely be wrong.

Solutions to the current AI writing problem are difficult. All I know is Lit's approach is too archaic, and could very well lead to the site becoming stagnant.

I really hope what you said about them using AI detectors as a sort of triage isn't true. Given what I've heard... meh, I don't know. They seem to not have many mods with the authority to approve stories though. Mine was held for almost a week before it was given the simple AI detected rejection. It's possible they just went through their backlog and purged a ton by handing out generic AI rejections for anything with even remote traces of potential AI use.
 
Solutions to the current AI writing problem are difficult. All I know is Lit's approach is too archaic, and could very well lead to the site becoming stagnant.
Your use of the word 'archaic' is interesting, it is a measure of the rate of change in the modern world.

I've been on AH for about 5 years. Early on, I submitted some images to accompany a story which included an express credit to the AI site - Creative Commons and all that. It was rejected on the grounds that it contained the name, not a link, of an outside website. I made appropriate changes, simply stating 'This is not a real person'. After a long wait, it was rejected again on the ground that I could not claim copyright in the images as required by Lit rules. No mention was made about the use of AI expressly.

Now Lit is expressly a 'No AI' domain and the rest of the world has moved on around it. Should I live so long, I look forward to what the next 5 years will render 'archaic'.
 
As far as anybody knows, there's only one person doing story approvals.
That's just not possible. It takes someone reading at a normal pace 5-6 hours to read 100k words. Doesn't Lit post hundreds of stories every day? Assuming at least 1k words per story...

So, either whoever is doing story approvals spends every waking hour with little to no sleep reading stories, or they themselves make extensive use of AI.
 
That's just not possible. It takes someone reading at a normal pace 5-6 hours to read 100k words. Doesn't Lit post hundreds of stories every day? Assuming at least 1k words per story...

So, either whoever is doing story approvals spends every waking hour with little to no sleep reading stories, or they themselves make extensive use of AI.

I think you missed the "as far as anyone knows" part.

Nobody really knows. Most of us quit speculating long ago, because doing so is fairly useless.
 
I think you missed the "as far as anyone knows" part.

Nobody really knows. Most of us quit speculating long ago, because doing so is fairly useless.
Ya, I was just following the hypothetical.

I'm still pretty new to the site, but do the owners hate transparency that much? And are they so concerned with control that they'd compromise the process? Backlogs nearly a week long publishing free content is something most would consider unacceptable, when they could shorten it to a fraction of that length with more volunteer mods.
 
That's just not possible.

Okay. So. Why does it make any difference? And why do you think your continuing harangues here are going to make any difference? Literotica does not accept AI-written, edited, or modified contributions now or in the foreseeable. There's an incredibly easy solution for you: find another site which does.
 
Status
Not open for further replies.
Back
Top