Resubmitted My Story After It Was Sent Back as AI Work — Seeking Clarity & Support

[XerXesXu]

I put that passage through a widely used AI detector and it returned a probability that it may be AI at 100%. It specifically mentioned the probability of expressions that would appear commonly in an LLM's training material. I didn't get a pay for analysis, but I can see many word pairings and triplets that are common, almost cliché. I can only guess, but when you learnt ESL, you may have also been trained to using high frequency constructions, words, and expressions in the expectation that those would be what you most commonly encounter.

 

[AwkwardlySet]

That's the issue here, I think. The extract is overwritten, one could highlight nearly every sentence to remove unnecessary words - which is a hallmark of AI content. @Five_Inch_Heels and @RagnarHairyBreeches highlight a sample of illogical or overblown sentence constructs, but there are many more. Not just a few, as you've suggested.

I do feel for ESL writers, not familiar enough with vernacular to know when to wind it back, but unfortunately this extract doesn't wind it back. I'm not surprised it's triggered whatever it is the site uses. Which might be Laurel's quick human eye scan, following on from content flagged by a bot, I don't know.

My recommendation to @Anjalisinha would be to (somehow) find a ruthless editor, someone not afraid to take the purple out of the prose, and challenge every sentence. A horrible prospect, for sure, not something that's easy to contemplate. I think the issue here is ESL, not AI, but the problem is, they can read the same.

You surely see the absurdity of what you are saying, EB? That every fresh author needs to first master writing before publishing on Literotica? This is exactly the place where amateurs are supposed to post their stories. And I judge the excerpt to be easily above Literotica-average.

When I first started publishing here, my first chapters were overwritten as well. I had a tendency to overexplain everything, along with mixing tenses, using confusing pronouns, you name it. And no, I disagree that the text reads as AI-generated. I've read a couple of stories on other sites that allow AI content, and this doesn't look the same to me. I am hardly an expert, but no one here is.

Take a look at this sentence, for example.

To show I meant it my comfort, my playfulness I gently placed a hand on each of their laps.

This isn't the type of mistake the AI would make. I don't think anyone can make sense of this. This screams amateur second language to me.

 

[Five_Inch_Heels]

You surely see the absurdity of what you are saying, EB? That every fresh author needs to first master writing before publishing on Literotica? This is exactly the place where amateurs are supposed to post their stories.

Master writing? No.

Have enough of a basic understanding of conversational English to make stories readable? Yes.

Note, there are sections for other languages for those who don't understand English well.

 

[AwkwardlySet]

Master writing? No.

Have enough of a basic understanding of conversational English to make stories readable? Yes.

Note, there are sections for other languages for those who don't understand English well.

I've seen plenty of stories with considerably worse English. Some of them are quite recent even.
I do partly agree with you, though. With such an enormous influx of daily stories, Lit has every reason to raise the standards for a story to be published. But keep in mind that this story hasn't been rejected because it wasn't well-written enough.

 

[Five_Inch_Heels]

I've seen plenty of stories with considerably worse English. Some of them are quite recent even.

You should see some of my early efforts.

 

[ElectricBlue]

This isn't the type of mistake the AI would make. I don't think anyone can make sense of this. This screams amateur second language to me.

Which is precisely what I said. The issue here is ESL, not AI. Problem is, the effect can be the same.

 

[AwkwardlySet]

I put that passage through a widely used AI detector and it returned a probability that it may be AI at 100%. It specifically mentioned the probability of expressions that would appear commonly in an LLM's training material. I didn't get a pay for analysis, but I can see many word pairings and triplets that are common, almost cliché. I can only guess, but when you learnt ESL, you may have also been trained to using high frequency constructions, words, and expressions in the expectation that those would be what you most commonly encounter.

There were also dozens of cases of rejected stories that went through free online AI-detection tools and came out as fully human-written. It's certainly not fair to look at things that way. Online tools saying it's AI-written is proof of AI-generated content, but online tools saying it's human-written isn't proof of human-generated content. There is no winning in that setup.

 

[Five_Inch_Heels]

Never trust a Bot to check for a Bot.

 

[AwkwardlySet]

Which is precisely what I said. The issue here is ESL, not AI. Problem is, the effect can be the same.

Indeed, and then you proceeded to speculate how Laurel could have seen it as AI-generated. Then you followed it with advice to find an editor who will completely overhaul the O.P.'s writing. You surely see this is a no-win scenario for the O.P.?

 

[XerXesXu]

There were also dozens of cases of rejected stories that went through free online AI-detection tools and came out as fully human-written. It's certainly not fair to look at things that way. Online tools saying it's AI-written is proof of AI-generated content, but online tools saying it's human-written isn't proof of human-generated content. There is no winning in that setup.

There are two characteristics of AI which AI detectors can pick up. The first is perplexity, the statistical chance of one word following another. The base model is trained like this on huge quantities of all qualities, but mostly, quite poor. Poor quality writing will not distinguish human from AI. The second is style, in the sense of 'style guide'. The garbage produced by the base model is fine-tuned, rewriting as trained by humans using a style guide. It's inevitable that there's an overlap between human and AI writing. Some humans will write like AI. The only advice they can be given is to write less predictability.

AI detectors may vary, as human opinions may also vary, but in the present case it's difficult to disagree with the detector. Too many high frequency couplets and triplets.

 

[ElectricBlue]

Indeed, and then you proceeded to speculate how Laurel could have seen it as AI-generated.

It's been rejected for suspected use of AI, so there's no speculation at all.

Then you followed it with advice to find an editor who will completely overhaul the O.P.'s writing. You surely see this is a no-win scenario for the O.P.?

Yes, it is, and where have I said it isn't? You clearly missed the use of the word "unfortunately".

I've not said the rejection is right, I've repeatedly said it's unfortunate, and I've made a suggestion for the OP to critically edit their story, or to find someone who can. Three years ago, before AI rejections became the thing, I'd have said exactly the same thing, and folk would have nodded, yep, that's good positive writers' advice. Where's the difference?

 

[Five_Inch_Heels]

I've not said the rejection is right,

The rejection is correct. The snippet as posted is unreadable and needs major rework, either by the OP or an editor with a lot of time on their hands.

 

[Anjalisinha]

That's the issue here, I think. The extract is overwritten, one could highlight nearly every sentence to remove unnecessary words - which is a hallmark of AI content. @Five_Inch_Heels and @RagnarHairyBreeches highlight a sample of illogical or overblown sentence constructs, but there are many more. Not just a few, as you've suggested.

I do feel for ESL writers, not familiar enough with vernacular to know when to wind it back, but unfortunately this extract doesn't wind it back. I'm not surprised it's triggered whatever it is the site uses. Which might be Laurel's quick human eye scan, following on from content flagged by a bot, I don't know.

My recommendation to @Anjalisinha would be to (somehow) find a ruthless editor, someone not afraid to take the purple out of the prose, and challenge every sentence. A horrible prospect, for sure, not something that's easy to contemplate. I think the issue here is ESL, not AI, but the problem is, they can read the same.

Am willing to work with editors
But where to find them.
No one replies from LIT VE list.
Or I didn’t get a reply from whom all I messaged.

 

[AwkwardlySet]

It's been rejected for suspected use of AI, so there's no speculation at all.

Yes, it is, and where have I said it isn't? You clearly missed the use of the word "unfortunately".

I've not said the rejection is right, I've repeatedly said it's unfortunate, and I've made a suggestion for the OP to critically edit their story, or to find someone who can. Three years ago, before AI rejections became the thing, I'd have said exactly the same thing, and folk would have nodded, yep, that's good positive writers' advice. Where's the difference?

I meant that you are speculating what could have made Laurel flag the story as AI.
I agree that ESL is the problem. It's the problem I had to deal with, too. But it seems Laurel treats ESL and AI the same, and that leaves the author with practically zero options. Solutions like "Get a knowledgeable editor who is willing to rewrite your story" or "upgrade your English skills by two tiers" aren't really helping the author. I know you are trying to help the O.P. the best you can, but you can see how absurd the O.P.'s options are.

My whole point is that the system should be designed in such a way as to allow honest authors to prove they are innocent of using AI. What Laurel has created is pure chaos instead, and a situation where we AH-ers have to deal with daily complaints and offer the same advice for the hundredth time, and witness borderline bullying from some of the people here who can't wait to point out the fallacies in rejected work, and usually hurry to justify Laurel's actions.
It's cult mentality, and it pisses me off. Either help the author with advice or action, or prove the obvious use of AI in their work. Both are fine. But this ridiculous sniping and tearing apart of someone's work, and here I don't mean you personally, is bullshit.

 

[Tomh1966]

Am willing to work with editors
But where to find them.
No one replies from LIT VE list.
Or I didn’t get a reply from whom all I messaged.

Ah the problem. Getting an editor. There are a few well known, but then tend to have well established connections to certain authors. yes, there are cliques in Lit.

A snippet of your snippet

“You travel alone often?” the younger one asked.

I glanced around dramatically, then whispered like I was confessing a secret.

Two observations:
1) Stilted as heck.
“You travel alone often?” the younger one asked.
more naturally would read
The younger one asked, "You travel alone often?"

I glanced around dramatically, then whispered like I was confessing a secret.
What I think you mean.
I looked around anxiously, then whispered like I was making a confession, "etc that next line here

Doing this line by line will be a LARGE job for an editor.

2) You seem to use words that are ALMOST synonyms quite a bit, but end up being wrong word choices. IE glanced for looked, confessing a secret for making a confession.

Sorry if this comes off mean, but that is how I see it.
In the end, editors are overwhelmed with requests, and even authors with 20+ works get turned down and/or told it will take months.

NEVER EVER copy paste a single word from Grammarly or other AI lite tool. They tend to put in overlarge and overflowery words.

 

[MillieDynamite]

First, @Anjalisinha, your writing isn't bad at all. But there are a few things that any AI checker might be triggered by.

“You travel alone often?” the younger one asked.

I glanced around dramatically, then whispered like I was confessing a secret.
“Only when I need to... explore myself. It’s easier to go deep when there’s no one watching.”

The older doctor gave a nervous chuckle.
“But in a crowded train, you don’t really get the privacy to go… deep in thought, you meant?”

The use of adverbs in what amounts to a dialogue tag, and the use of dialogue tags preceding the quotes and having them in what appears to be the same paragraph, would cause AI checkers to tag those lines.

I turned toward him, my eyes wide with playful horror.
“Oh no no, don’t get the wrong idea. I meant deep in conversations.”

I paused, then added softly, “I enjoy going deep with strangers ... just mentally, of course.”

He cleared his throat, clearly not buying the innocence but unsure how to call it out.

“By no one watching, I just meant away from family,”
“They’re always checking on me. Asking questions. Watching my every move. Here, I can be... more open. Speak without filters & do what I want”

To show I meant it my comfort, my playfulness I gently placed a hand on each of their laps. A soft pat on their thighs. Not sexual. Not yet. Just reassurance. A teasing reminder that I wasn’t scared of closeness. Not anymore.

Playful horror is an odd comparison in English, and it might trigger an AI checker to go ding, ding, ding. The repetition of variants of the same word so close together might trigger AI checkers—playful and playfulness.

There is nothing significant that screams 'AI,' but if your human rating is below some unknown percentage, between 91 and 96 percent is my guess, you'll get rejected.

 

[Bramblethorn]

First of all, thank you for such a detailed and structured reply.
Yes, English isn’t my first language.

I'd also guess that the variety of English you know best is Indian English. There are some differences between Indian English and the varieties commonly spoken in the UK and USA, and I wonder if those might lead to higher error rates for "AI detectors" in assessing Indian English writing.

 

[Five_Inch_Heels]

Maybe she can have both happen. Get tossed out of the McMansion and into trailer park working at the diner.

It's in those two places that she meets the people she used to look down her nose at and finds out she has to depend on them. Hey, if she has to start hooking .... selling her ass to save her ass ...

 

[RagnarHairyBreeches]

There is a solution to your problem. I don't think getting an editor is one of them. Sure, it would work, but you'd be paying them a fortune. The story would improve but your language skills would not.

Train yourself in English usage.

Read more. Go to your local library and read as many novels as you can. Different authors. Study the way they write. Listen to the conversations of English speakers (don't be rude and eavesdrop). You might be able to enroll in an evening class, or join a support group.

My only language is English. I hated grammar when I was at school and I still don't understand all the rules, but I think I have a good grasp of it to get me by. So, I'd advise against spending time reading English grammar texts. If you know anyone whose only language is English, then speak to them and if they're willing, ask them to criticize your spoken English on a day-by-day basis. And listen to them and remember what they say, and how they say it. If the best your English skills become is how people talk in a pub, or how people talk in the mall, then you will have improved.

You watch TV. Movies. The evening news service. Listen carefully to how they use language and try to learn from them. I'm suggesting that because it's easy and won't cost you anything. Perhaps it's not the best solution but it may help you. Then, when you feel more confident you can try for a book on grammar. Though, many of English grammar rules are stupid.

Good luck!

 

[erotica_n_s]

I'm sorry if this question has been asked a million times before - but is any AI use permitted? I'm not intending to enter my work into a contest or any such thing, I'm not a professional writer, I've had a few story ideas sitting around for several years, never had the time to write them out as I would want to, but now, thanks to AI, I thought I might give it a go - not by using AI to write the whole story, but using it to edit and polish my work, etc. Is that not allowed on Literotica?

If not here at Literotica, then... is there anywhere else online I could post such work - not entirely AI generated - far from it, but work that used a certain amount of AI in its preparation (and makes that clear)?

 

[Bamagan]

I'm sorry if this question has been asked a million times before - but is any AI use permitted? I'm not intending to enter my work into a contest or any such thing, I'm not a professional writer, I've had a few story ideas sitting around for several years, never had the time to write them out as I would want to, but now, thanks to AI, I thought I might give it a go - not by using AI to write the whole story, but using it to edit and polish my work, etc. Is that not allowed on Literotica?

If not here at Literotica, then... is there anywhere else online I could post such work - not entirely AI generated - far from it, but work that used a certain amount of AI in its preparation (and makes that clear)?

Using it as a 'polisher' is not allowed on Literotica. Very minimal use such as identifying spelling and grammar errors is likely to be okay, but non-AI software that does that just fine has been around for many years now. The newer stuff that is 'powered' by LLMs does more than highlight potential errors, and letting it rewrite sentences to conform to its 'ideals' will likely trigger a rejection from the site.

 

[erotica_n_s]

If not here at Literotica, then... is there anywhere else online I could post such work - not entirely AI generated - far from it, but work that used a certain amount of AI in its preparation (and makes that clear)?

 

[Bamagan]

If not here at Literotica, then... is there anywhere else online I could post such work - not entirely AI generated - far from it, but work that used a certain amount of AI in its preparation (and makes that clear)?

Probably. I don't know of any off the top of my head, but I imagine you can try looking it up.

 

[RagnarHairyBreeches]

I don't think AI would be worth using, even if it were permitted. Have you seen those AI generated images that show people with six fingers or arms popping through someone's body. It's all very nice in principle, but a load of rubbish.

There was a US legal case that got stopped in its tracks. The lawyer used one of those software packages to help prepare their case. To prepare their argument. It was working well until until a law was cited that didn't exist. The AI had made it up. The judge pulled the lawyer up quick smart. The software does stupid things. And that must have been a really dumb/lazy lawyer.

I use Grammarly. I know the difference between it's and its but my typing fingers keep getting the apostrophe wrong. Grammarly is good at picking up stuff like that. I use the free version. The software keeps trying to encourage me to upgrade to the paid version by tempting me by showing me improvements. I sometimes click on the suggestions. And, you know what, most of the time I like my version better. Sometimes the suggested change alters the entire meaning of the sentence.

These AI tools are not as good as you'd like to hope for. They get a lot of stuff right, but they get a lot of stuff wrong too. They're best avoided, for what you're thinking, in my opinion. Just use them as tool. To help with your work. Not to create it. Actually, now I'm thinking of it, if you use MS Word, one of the best things on there is Review > Read Aloud. Once your text is done, or nearly done, let the software read back what you've written. It's amazing the amount of issues you can pickup by hearing it read to you.

 

[erotica_n_s]

Yes, I know... but the thing is, I'm not a professional writer, I don't have the time or the creative patience to sit down and write out my story in full the way I'd like to... I've had some vanilla story ideas for years that I would love to see written out in full, but it just hasn't happened - I don't have the time. So in recent days, I've been experimenting with AI, giving it my prompts and working with it 'collaboratively', it's been an interesting process. I'm not aiming to pass it off as my own work, if anything, the intention was to present the finished work, and then, separately, to demonstrate how I went through the writing process, and where and how I used AI. If I can't do that on Literotica, it's a disappointment, but I'll have to find somewhere else to publish it.