Story rejected for AI: Now what?

[millamomud]

If your fiction writing is mistaken for AI slop, that might be a sign that your style isn't suited to fiction writing. People have had their stories rejected for "not the style we're looking for" forever, here and by trad publishing houses. If Lit decides your story isn't up to the site's standards, it's their right not to publish it.

Just because Literoticas misguided detector thinks you sound like an AI doesn't mean you write like slop. I think you give the detector too much credit.

 

[Dark_Logan_]

What feedback would you like, that would not constitute advice on how to disguise your use of AI?

Knives out with the accusations

I presume that you’ve assessed every one of the tales I’ve penned in the five years of stories submitted here… or at least run them through Grok to assess them for likely AI content

 

[Dark_Logan_]

I don't think the problem is authors using AI to proofread stories. That's really no different than having a human editor, but is fraught with the same problems.

The problem is authors automatically using the changes the proofreader suggests. The key is to use very simple proofreading, meaning spelling and punctuation only, and not accepting any suggested change as the Gospel According to (fill in the name of the software used). AI must be taught how to do things and in the case of proofreading, the lessons come from other written works. Depending upon how AI was taught, those written works may include fiction, scientific papers, business articles, and almost any other type of writing including erotica, along with some literary "rules" that may vary depending upon the opinion of the person who wrote the rules.

Look at every suggested change and decide if the change sounds like the way real people speak every day, not how they speak when giving a presentation at a board meeting.

I’m sorry to make this about me… but it’s the only proof I have on matters.

A story submitted around the 20th October for Halloween with (honest to god) no use of AI and nothing other than Microsoft Word for a spellcheck still pending publishing in early January suggests otherwise

With 107 previous trouble free submissions I’ve tried every theory here for resubmission, reached out multiple times following two rejections, one for suspected AI and then one for the suspected use of AI based tools, and ranted into this void countless times. I’ve all but rewritten many aspects of the story and yet still the outcome is the same with not one iota of inclination from the individuals that host the dinner party as to how I get a seat at the table

Fair to say I’ve lost my appetite as a consequence

 

[BobbyBrandt]

Just because Literoticas misguided detector thinks you sound like an AI doesn't mean you write like slop. I think you give the detector too much credit.

You call it "misguided' yet thousands of other writers find it to work just fine.

Is it any wonder that the site isn't making changes to their AI detection procedure a top priority? A small percentage of submissions get rejected for suspected AI yet those that receive those rejections expect everyone else to acquiesce to their claims of false detection.

Some may be false, but there is plenty of proven advice here on how to remedy those situations for those who aren't too stubborn to do do.

 

[BobbyBrandt]

I’m sorry to make this about me… but it’s the only proof I have on matters.

A story submitted around the 20th October for Halloween with (honest to god) no use of AI and nothing other than Microsoft Word for a spellcheck still pending publishing in early January suggests otherwise

With 107 previous trouble free submissions I’ve tried every theory here for resubmission, reached out multiple times following two rejections, one for suspected AI and then one for the suspected use of AI based tools, and ranted into this void countless times. I’ve all but rewritten many aspects of the story and yet still the outcome is the same with not one iota of inclination from the individuals that host the dinner party as to how I get a seat at the table

Fair to say I’ve lost my appetite as a consequence

Simply because you can't see the problem in what you submitted doesn't mean that isn't really there. Maybe ask someone else to take a look at it to see if they see something that you don't.

Your ability to publish successfully in the past demonstrates that the publishing procedures work for you most of the time. There is no evidence that there has been any change to the AI detection process used by the site in the last year, but it's possible.

 

[AwkwardlySet]

Yeah, this is devolving into another one of the AH-specials.

To all of you who received a rejection for AI, I suggest considering what @ShelbyDawn57 wrote. A good part of that post contains good general writing advice, and the rest you should see as the necessary evil in order to be approved. Unlike most others, that post is also delivered with a kind and helpful attitude. It's good advice altogether.
@ShelbyDawn57 for moderator.

Most of the rest is the typical AH pretentious bullshit that states that if Literotica's second-rate detection algorithm has flagged your work, it must mean that you either cheated by using AI to write your story, or your writing style sucks anyway and reads like AI-slop.
Some of those attitudes stem from the cultist attitude some AH-ers have towards Laurel and Lit, and some stem from pure pretentiousness and a sense of superiority. Some come from both, I suppose.

Anyway, if that were actually so, if Laurel's algorithm actually detected bad writing, then the overwhelming mass of garbage stories wouldn't get published every day. You just need to browse through the new story list and see for yourself what kind of rubbish gets published. And that's all fine. The site doesn't actually have any writing standards or requirements past some basic punctuation and structuring.

But then, the cultists should stop demeaning the authors who get rejected for AI, because writing badly CLEARLY isn't what triggers the rejection.

 

[StillStunned]

Yeah, this is devolving into another one of the AH-specials.

To all of you who received a rejection for AI, I suggest considering what @ShelbyDawn57 wrote. A good part of that post contains good general writing advice, and the rest you should see as the necessary evil in order to be approved. Unlike most others, that post is also delivered with a kind and helpful attitude. It's good advice altogether.
@ShelbyDawn57 for moderator.

Most of the rest is the typical AH pretentious bullshit that states that if Literotica's second-rate detection algorithm has flagged your work, it must mean that you either cheated by using AI to write your story, or your writing style sucks anyway and reads like AI-slop.
Some of those attitudes stem from the cultist attitude some AH-ers have towards Laurel and Lit, and some stem from pure pretentiousness and a sense of superiority. Some come from both, I suppose.

Anyway, if that were actually so, if Laurel's algorithm actually detected bad writing, then the overwhelming mass of garbage stories wouldn't get published every day. You just need to browse through the new story list and see for yourself what kind of rubbish gets published. And that's all fine. The site doesn't actually have any writing standards or requirements past some basic punctuation and structuring.

But then, the cultists should stop demeaning the authors who get rejected for AI, because writing badly CLEARLY isn't what triggers the rejection.

So Shelby says, "Here are some tips to improve your writing. They might help you less likely to be mistaken for AI." Which is great, because they're useful tips. And you applaud them.

And someone replies that they shouldn't have to improve their writing just to be published here. Others explain that they might have to, and no-one has an inalienable right to be published if their style isn't what Lit wants. But you say that's bad? Even though it's basically supporting what Shelby said?

 

[RejectReality]

Yeah, this is devolving into another one of the AH-specials.

To all of you who received a rejection for AI, I suggest considering what @ShelbyDawn57 wrote. A good part of that post contains good general writing advice, and the rest you should see as the necessary evil in order to be approved. Unlike most others, that post is also delivered with a kind and helpful attitude. It's good advice altogether.
@ShelbyDawn57 for moderator.

Most of the rest is the typical AH pretentious bullshit that states that if Literotica's second-rate detection algorithm has flagged your work, it must mean that you either cheated by using AI to write your story, or your writing style sucks anyway and reads like AI-slop.
Some of those attitudes stem from the cultist attitude some AH-ers have towards Laurel and Lit, and some stem from pure pretentiousness and a sense of superiority. Some come from both, I suppose.

Anyway, if that were actually so, if Laurel's algorithm actually detected bad writing, then the overwhelming mass of garbage stories wouldn't get published every day. You just need to browse through the new story list and see for yourself what kind of rubbish gets published. And that's all fine. The site doesn't actually have any writing standards or requirements past some basic punctuation and structuring.

But then, the cultists should stop demeaning the authors who get rejected for AI, because writing badly CLEARLY isn't what triggers the rejection.

Got to agree. I'm the last one to tell you to change your style for the sake of Lit's hallucinating clanker, but there's plenty of good advice for making your writing more interesting in that post. Don't think it will do diddly squat for anyone who genuinely didn't let AI rewrite anything for them and got rejected for it anyway.

If bland, sterile writing was what was triggering it, 3/4 of the new authors would be getting rejected, and that's not what's happening.*

* 3/4 of those who do get published before anyone goes on about not knowing how many are rejected.

 

[millamomud]

Yeah, this is devolving into another one of the AH-specials.

To all of you who received a rejection for AI, I suggest considering what @ShelbyDawn57 wrote. A good part of that post contains good general writing advice, and the rest you should see as the necessary evil in order to be approved. Unlike most others, that post is also delivered with a kind and helpful attitude. It's good advice altogether.
@ShelbyDawn57 for moderator.

Most of the rest is the typical AH pretentious bullshit that states that if Literotica's second-rate detection algorithm has flagged your work, it must mean that you either cheated by using AI to write your story, or your writing style sucks anyway and reads like AI-slop.
Some of those attitudes stem from the cultist attitude some AH-ers have towards Laurel and Lit, and some stem from pure pretentiousness and a sense of superiority. Some come from both, I suppose.

Anyway, if that were actually so, if Laurel's algorithm actually detected bad writing, then the overwhelming mass of garbage stories wouldn't get published every day. You just need to browse through the new story list and see for yourself what kind of rubbish gets published. And that's all fine. The site doesn't actually have any writing standards or requirements past some basic punctuation and structuring.

But then, the cultists should stop demeaning the authors who get rejected for AI, because writing badly CLEARLY isn't what triggers the rejection.

Well said

 

[BobbyBrandt]

But then, the cultists should stop demeaning the authors who get rejected for AI, because writing badly CLEARLY isn't what triggers the rejection.

There is enough bad writing here that doesn't get flagged for AI to validate your point. In fact, as some have pointed out, bad writing is less likely to trigger an AI rejection due to the fact that it IS bad writing.

What triggers an AI rejection here? It's anyone's guess.

What is clear is that not everyone is getting tripped up in AI rejections, with hundreds of stories being published each month. My position is that blaming the detection process for 100% of the rejections is plain idiotic. Several people have posted their experiences where identifying a possible AI element or style and then correcting it resulted in the story getting published.

No one is disputing the possibility of false AI detection. What is being disputed are the conspiracy theories and hysteria over the few that get rejected compared to the thousands that get approved. That, and the stubborn refusal of many writers to accept proven advice that might help them overcome the challenges that they have faced.

Whether the cause for an AI rejection was self-inflicted or systemic to the site, either try to correct the work where necessary or don't. Bitching about it is pointless.

 

[Gamblnluck]

I spent a month working on a new story and I'm really proud of it. I was so excited to post it, but it was rejected for AI. I resubmitted with a note explaining that I didn't use AI (only the basic spellcheck in MS Word), but it is still rejected.

Was that basic spell check, grammarly? I've heard that program sets of AI detectors. A college prof friend of mine told me it sets off their alerts as well.

If so, I'd suggest doing sort of a re-write. Go through it line by line, changing the syntax a bit. Then re-submit.

 

[AwkwardMD]

As much as I default to Lit's side on AI rejections, I'm the first to admit that false positives are possible. That being said, it's not a coincidence that two out of the three loudest voices (in this thread, anyway) complaining that the AI detector is faulty use AI-generated pfps. Their "I would never" or "I didn't" requires a footnote.

 

[Dark_Logan_]

Ah an awkward MD cometh with her/his opinion just for some attention

Behold she/he who must always have a vague/ambiguous opinion on how this site is run

Whilst providing no tangible help or assistance but cast aspiration that profile pictures must now also be real to be taken serious

Thread complete, can an Admin lock this for further comment please



I’m off to get Grok to put a bikini on my profile pic before it writes me a few more chapters for my upcoming autobiography titled “I did it my way” …will only be available as an ebook on Amazon

 

[AwkwardMD]

Ah an awkward MD cometh with her/his opinion just for some attention

Behold she/he who must always have a vague/ambiguous opinion on how this site is run

Whilst providing no tangible help or assistance but cast aspiration that profile pictures must now also be real to be taken serious

Thread complete, can an Admin lock this for further comment please



I’m off to get Grok to put a bikini on my profile pic before it writes me a few more chapters for my upcoming autobiography titled “I did it my way” …will only be available as an ebook on Amazon

Okay

 

[Five_Inch_Heels]

So Shelby says, "Here are some tips to improve your writing. They might help you less likely to be mistaken for AI." Which is great, because they're useful tips. And you applaud them.

And someone replies that they shouldn't have to improve their writing just to be published here. Others explain that they might have to, and no-one has an inalienable right to be published if their style isn't what Lit wants. But you say that's bad? Even though it's basically supporting what Shelby said?

Helping the OP is fine.

Interpersonal bickering is not.

 

[TheWritingGroup]

Knives out with the accusations

I presume that you’ve assessed every one of the tales I’ve penned in the five years of stories submitted here… or at least run them through Grok to assess them for likely AI content

What? I wrote:

What feedback would you like, that would not constitute advice on how to disguise your use of AI?

I meant, how could Literotica explain exactly how their detector works without telling AI-text-stealers how to sneak works past it? Nowhere in there do I say, "And you are one of the AI-assisted text thieves." Because I did not mean that.

--Annie

 

[IWroteThis]

Apparently, the trick is to state up front that you used AI. From a story published today:

"I used AI tools to structure this story but the story plot is entirely mine."

Yes, the author really put that in their Foreword.

https://www.literotica.com/s/a-steamy-night-at-delhi-with-ex

 

[AwkwardlySet]

Apparently, the trick is to state up front that you used AI. From a story published today:

"I used AI tools to structure this story but the story plot is entirely mine."

Yes, the author really put that in their Foreword.

https://www.literotica.com/s/a-steamy-night-at-delhi-with-ex

Oh my god, this is hilarious.
Well, I suppose if we concluded that there are false positives, there are bound to be false negatives too. But either way, this proves that no human looks at our stories, or at the very least, that no human looks at the stories that didn't get flagged by the algorithm. There's no way that Laurel went through this story yet didn't bother with the foreword.

 

[old_prof]

Oh my god, this is hilarious.
Well, I suppose if we concluded that there are false positives, there are bound to be false negatives too. But either way, this proves that no human looks at our stories, or at the very least, that no human looks at the stories that didn't get flagged by the algorithm. There's no way that Laurel went through this story yet didn't bother with the foreword.

You do realize what skim reading means, right?

 

[AwkwardlySet]

You do realize what skim reading means, right?

You're insane if you believe that she skim-reads 170+ stories every day. More than that, if that were the method of approval that all stories had to go through, then we wouldn't have all these stories sitting in pending for weeks and months. She'd just approve or reject them after that skim-read. Your theory makes no sense.

 

[old_prof]

You're insane if you believe that she skim-reads 170+ stories every day. More than that, if that were the method of approval that all stories had to go through, then we wouldn't have all these stories sitting in pending for weeks and months. She'd just approve or reject them after that skim-read. Your theory makes no sense.

First off it's not that hard to ski read a few thousand pages a day. There are literally dozens of examples in the last few months that can only have been seen by a human looking at the story. You've never had even a vaguely coherent theory that explains those. You simply just ignore any information that disagrees with your cock-a-manie view of how things must work to feed your something.

I'm done interacting with you. Go find somewhere else you can piss to show how big your dick is.

 

[AwkwardlySet]

First off it's not that hard to ski read a few thousand pages a day. There are literally dozens of examples in the last few months that can only have been seen by a human looking at the story. You've never had even a vaguely coherent theory that explains those. You simply just ignore any information that disagrees with your cock-a-manie view of how things must work to feed your something.

I'm done interacting with you. Go find somewhere else you can piss to show how big your dick is.

Wow. Resorting to ad hominems has become so typical here. That's probably why you seem to fit so well in the AH.

But since we're there, let's summarize your relatively brief AH history. You went from the typical first-time writer full of enthusiasm, the same as we all were once, to the outright worshipper of the website fairly quickly. Going after those who criticized Lit and all that.

But then, the problem with the rendering of your story happened, and you went full ballistic in the thread you made, calling Manu all kinds of names and mincing no words about the website's capabilities and standards.
Then you edited your story and band-aided the problem, for yourself only, as the bug still persisted, and as far as I know, still persists in some form. And that was enough to make you switch back to the worshipper side.

Yeah, I'd pay more attention to your words if you weren't switching sides faster than a ping-pong ball.

 

[AwkwardMD]

I keep saying this, but I'm not sure it's sinking in. When two people say "I only use Grammarly to" (or whatever), there could be an ocean between the practical implementations.

If this story got through the detector, I believe that the author wrote it themselves. If they used an AI tool, they used it differently. Maybe they had a conversation with a chat bot, got some ideas, and then turned around and wrote the story themselves. That seems perfectly permissible and aligned with the author's admission in the foreword. Especially given that the author was up front about it, I see no reason to report it or demean it.

Laurel is not skimming submissions except on resubmits after a rejection. She never did that before (or if she did it was way in the past even when I was fresh faced), and she's unlikely to start that now that submissions have quadrupled.

 

[AwkwardlySet]

I keep saying this, but I'm not sure it's sinking in. When two people say "I only use Grammarly to" (or whatever), there could be an ocean between the practical implementations.

Absolutely. Authors should understand the implications of using Grammarly's suggestions to rewrite their sentences. That's using AI to write the story for you.

If this story got through the detector, I believe that the author wrote it themselves. If they used an AI tool, they used it differently. Maybe they had a conversation with a chat bot, got some ideas, and then turned around and wrote the story themselves. That seems perfectly permissible and aligned with the author's admission in the foreword. Especially given that the author was up front about it, I see no reason to report it or demean it.

I'm sure you know the principle of Occam's razor. Which do you think is more likely, that the author didn't use AI to generate any words but only to "structure" their story, whatever structuring meant for them, or that Laurel's algorithm isn't perfect? I mean, I can't figure out what I would even ask the AI if that didn't involve generating text or asking to generate plot ideas (as the author claims that plot is 100% theirs).

For me, the likelihood that Laurel's algorithm simply failed to flag is far greater, especially in the sense I already mentioned - the existence of false positives implies imperfection, which means that it's likely there are false negatives too.

Laurel is not skimming submissions except on resubmits after a rejection. She never did that before (or if she did it was way in the past even when I was fresh faced), and she's unlikely to start that now that submissions have quadrupled.

That's my opinion too. We are all guessing, but I find that view far more reasonable than the view that she skim-reads all stories.

 

[AwkwardMD]

Absolutely. Authors should understand the implications of using Grammarly's suggestions to rewrite their sentences. That's using AI to write the story for you.

It makes it really tricky to interface with new authors I don't know, who use all the same words while trying to state their case.

I'm sure you know the principle of Occam's razor. Which do you think is more likely, that the author didn't use AI to generate any words but only to "structure" their story, whatever structuring means for them, or that Laurel's algorithm isn't perfect? I mean, I can't figure out what I would even ask the AI if that doesn't involve generating text or asking to generate plot ideas (as the author claims that plot is 100% theirs).

For me, the likelihood that Laurel's algorithm simply failed to flag is far greater, especially in the sense I already mentioned - the existence of false positives implies imperfection, which means that it's likely there are false negatives too.

I used to be more diligent about periodically pointing out that I do not empirically know how Lit's AI Detector works. I believe that I have reverse engineered it, but there's no way for me to get under the hood of the website and find out for sure. That's not in my tool set. I don't know, but my (capital t) Theory continues to hold water with every rejection I come across, and that gives me enough confidence to continue to state "I know how it works." So, with that in mind...

I have always admitted the unfortunate possibility of false positives. Even from the very start, almost two years ago in private, I was saying that. I immediately understood how it was possible.

False negatives are not possible. They just aren't. (EDIT: my point, here, was for anyone reading this to take this entire section in total. It's not possible within my Theory.)

That's my opinion too. We are all guessing, but I find that view far more reasonable than the view that she skim-reads all stories.

The site's policy of "resubmit with a note" worked on one of my earliest submissions, and it was Laurel herself in 2014 saying "if you really believe it was in error, send it in again. I'll take a look at it." I had a 16 year old character in the story, but they were nowhere near the sex. Upon resubmission (without having changed anything), it was posted. This is the way it has always worked (in my time, anyway).