Do moderators use algorithms to reject stories?

[EmilyMiller]

Their work was probably rejected because it was too grammatically correct and hence, the algorithms marked it a AI.

This is BS.

This is not how AI detectors work and frankly it’s an insult to the authors of many perfectly grammatical works that get published here every day.

I’m kinda sick of hearing this. It’s not true.

 

[BaldMerkin]

But it isn't about quality.

It's about adhering to the rules.

None of the rules say a submitted story has to be "good enough."

I sometimes wish that the rules do. I've tried to read some stories on Lit and the grammatical and punctuation errors make it impossible to understand.

 

[EmilyMiller]

Melville's Moby Dick has been identified as being AI generated

Cite which AI detector did that.

 

[SmilingLez]

There is something wonderfully confident about declaring what an imperfect detection system can and cannot do while simultaneously insisting that rejected authors must be the ones mistaken.

Apparently, algorithms are fallible right up until the explanation requires admitting that they might actually be fallible.

 

[BaldMerkin]

This is BS.

This is not how AI detectors work and frankly it’s an insult to the authors of many perfectly grammatical works that get published here every day.

I’m kinda sick of hearing this. It’s not true.

What's BS? That Moby Dick was identified as written by AI or that it's too grammatically correct? Because I assure you that Moby Dick was definitely identified by AI detectors to be AI written. A University lecturer did the analysis, using the tools they use to detect assignments written by AI, and I work in IT. Do you?

 

[SmilingLez]

What's BS? That Moby Dick was identified as written by AI or that it's too grammatically correct? Because I assure you that Moby Dick was definitely identified by AI detectors to be AI written. A University lecturer did the analysis, using the tools they use to detect assignments written by AI, and I work in IT. Do you?

Here's the link you're referring to

https://www.linkedin.com/pulse/moby-dick-failed-ai-test-markus-brinsa-6kmrc?tl=en

 

[ChuckWolf]

That's not quite true. I've heard of authors who definitely do NOT use AI but have been accused of doing so. Their work was probably rejected because it was too grammatically correct and hence, the algorithms marked it a AI.

Melville's Moby Dick has been identified as being AI generated but it was written well before AI was invented. Using algorithms is fallible so a story shouldn't be rejected because it IS grammatically correct.

Right, but under my initial idea, the writers who would be rubber-stamped in publishing would be long time (10 years?) writers who have 40 stories AND 200 words written; I don't think most of these writers are likely to use AI.

I mean, you're always going to have special cases of course.

 

[EmilyMiller]

What's BS? That Moby Dick was identified as written by AI or that it's too grammatically correct? Because I assure you that Moby Dick was definitely identified by AI detectors to be AI written. A University lecturer did the analysis, using the tools they use to detect assignments written by AI, and I work in IT. Do you?

Please provide a link - please provide what tool they used.

It’s BS that grammatically correct text gets eroneously pinged as AI. It’s BS that introducing random errors is a defense against being pinged for AI.

Working in IT doesn’t guarantee an understanding of how LLMs work. Having a background in statistics is rather more useful.

Pangram thinks that this one section from the first 1,000 words might be AI, but with low confidence. It thinks the other 900 words are human written:

Whenever I find myself growing grim about the mouth; whenever it is a damp, drizzly November in my soul; whenever I find myself involuntarily pausing before coffin warehouses, and bringing up the rear of every funeral I meet; and especially whenever my hypos get such an upper hand of me, that it requires a strong moral principle to prevent me from deliberately stepping into the street, and methodically knocking people's hats off--then, I account it high time to get to sea as soon as I can. This is my substitute for pistol and ball.

 

[BaldMerkin]

This is BS.

This is not how AI detectors work and frankly it’s an insult to the authors of many perfectly grammatical works that get published here every day.

I’m kinda sick of hearing this. It’s not true.

I assure you that I personally know the author and they had to change their story, that was identified as being written by AI, in order for it to be published. They made changes to: spelling, grammar and punctuation. After those changes the story was accepted. Your statement of BS is BS.

I feel that your responses are getting close to abusive so please be more civil in your posts.

 

[BaldMerkin]

Please provide a link - please provide what tool they used.

It’s BS that grammatically correct text gets eroneously pinged as AI. It’s BS that introducing random errors is a defense against being pinged for AI.

Working in IT doesn’t guarantee an understanding of how LLMs work. Having a background in statistics is rather more useful.

Pangram thinks that this one section from the first 1,000 words might be AI, but with low confidence. It thinks the other 900 words are human written:

Someone else has provided you with a link to a similar article. The point is that AI detectors are NOT infallible. Trusting them to reject stories, without details for the rejection, is not a valid way of deciding whether a story should or shouldn't be detected.

 

[WeShallUnclench]

2. Be careful about stories that take place in current times involving politics, even international politics.

Wut

I mean I get if you post propaganda that falls on a side of the politial spectrum that advocates for ethnic cleansing and shit, or if you side with a certain someone who thinks their western borders were just a suggestion.
But a general "don't talk politics" would be... bad.
Well, I suppose specific campaign messages would also be inappropriate, in any case, that I could also understand.

Anyway, regarding automatized rejections - one of my stories was stuck in limbo for a bit, and I THINK it was because it featured a naked toddler. Looks like a classic case of, software sees a combination of words and springs to action.
Considering that it WAS ultimately approved, considering there was quite obviously no sexual context for that character in the story, I'd say either a better software took a look at it, or indeed Laurel checked it out.
Speculation, of course.

 

[ElectricBlue]

Wut

I mean I get if you post propaganda that falls on a side of the politial spectrum that advocates for ethnic cleansing and shit, or if you side with a certain someone who thinks their western borders were just a suggestion.
But a general "don't talk politics" would be... bad.
Well, I suppose specific campaign messages would also be inappropriate, in any case, that I could also understand.

The site's view is that readers want erotica as an escape from overt politics in story content. This stance tightened during the Trump - Clinton election. It's going to be far tighter now.

Anyway, regarding automatized rejections - one of my stories was stuck in limbo for a bit, and I THINK it was because it featured a naked toddler. Looks like a classic case of, software sees a combination of words and springs to action.

Correct. A naked toddler is definitely going to trigger a bot.

Considering that it WAS ultimately approved, considering there was quite obviously no sexual context for that character in the story, I'd say either a better software took a look at it, or indeed Laurel checked it out.
Speculation, of course.

Was it rejected, or just took a while to get processed?

 

[BobbyBrandt]

I'm not responsible for what bothers you.

I wasn't trying to make some revolutionary statement. She wants certain kinds of stories on this site, yes? Those don't include stories with underage sex, yes?

I impute no motive whatsoever to Laurel. I couldn't care less why she doesn't want these stories here. I am aware that she doesn't, and that's good enough for me. For all of us, in fact.

Frankly, I find "want" to be a very odd word for you and other posters to parse. Are you implying that she in fact WANTS underage stories here? Because that's the obvious rejoinder to what I wrote, and I see no evidence of it.

"Want" implies prudery to you, perhaps, but not to me. And I did not intend it that way. Not that it matters; the point I was making is that her motivation doesn't matter. Only her rules matter, surely.

"Want" has literally nothing to do with the content guidelines here.

Based upon the types of stories (and substandard quality of some), I think that Laurel has demonstrated that she holds little bias where story submissions are concerned. As long as they pass the "industry standards" for content used on this site and most others, the story gets published.

 

[EmilyMiller]

I feel that your responses are getting close to abusive so please be more civil in your posts.

Funny how that happens. Easy remedy.

 

[SmilingLez]

I feel that your responses are getting close to abusive so please be more civil in your posts.

In an anonymous online forum on a for-profit commercial porn site, a few bullies are going to slip through the cracks. Best bet (YMMV) report them. Hypothetically, they are interesting to study like a bug under a microscope.

 

[ronde]

Wut

I mean I get if you post propaganda that falls on a side of the politial spectrum that advocates for ethnic cleansing and shit, or if you side with a certain someone who thinks their western borders were just a suggestion.
But a general "don't talk politics" would be... bad.
Well, I suppose specific campaign messages would also be inappropriate, in any case, that I could also understand.

Anyway, regarding automatized rejections - one of my stories was stuck in limbo for a bit, and I THINK it was because it featured a naked toddler. Looks like a classic case of, software sees a combination of words and springs to action.
Considering that it WAS ultimately approved, considering there was quite obviously no sexual context for that character in the story, I'd say either a better software took a look at it, or indeed Laurel checked it out.
Speculation, of course.

The story in question referenced the current politics in a country that was usually in the news of the day. The political climate was just used to explain the plot of the story, and there was no justification of any ideology. I re-wrote the same story but placed it in the past and it went through with no problems. I have written other stories using politics to set the stage for the story and as long as the time frame is the past or even the future, I haven't had any problems.

I think the problem with current politics in a story is that some would use Literotica to advance their particular ideology rather than just telling a story. If you take a look at the politics forum, you'll see that this happens in nearly every post, and many of those posts are unencumbered by facts other than the author's opinion. That isn't story telling. It's as you say, just propaganda.

 

[WeShallUnclench]

Correct. A naked toddler is definitely going to trigger a bot.

Was it rejected, or just took a while to get processed?

It just got stuck in pending for a week

 

[BaldMerkin]

My friend, a university lecturer, passed on some investigations to check for AI generated content. A summary of the results:

“Pride and Prejudice” by Jane Austen Your Text is AI/GPT Generated 100% by ZeroGPT.

“Moby-Dick” by Herman Melville Your Text is AI/GPT Generated 100% by ZeroGPT.

Walden — Henry David Thoreau Your Text is AI/GPT Generated 100% by Grammarly.

Dracula — Bram Stoker Your Text is AI/GPT Generated 100% by Grammarly.

All those novels were written before AI.

Turnitin, which many lecturers use to test submitted assignments, suggests that it should NOT be used as a single determinant of AI use to generate text, usually student submissions.

The point is that, any tool that returns a result should be assessed by a human to make the final decision.

I just thought that this might be interesting to some people.

 

[ElectricBlue]

I just thought that this might be interesting to some people.

The use of AI and false positives has been a regular topic in the AH for the last two years. Many of us (most of us?) are well and truly over it now.

Whatever and however the site does what it does, the number of threads on this topic seems to be reducing, which means something is sorta kinda working. That's good enough for me.