How to Avoid AI Rejection

[Priscilla_June]

This is a good question. I'd like to hear from a) experienced Lit authors if they've encountered an AI rejection, and b) new Lit authors who have NOT encountered this problem.

I would say given I joined in 2022 and only have 22 stories I am in between? I’ve never once been rejected for AI.

 

[AwkwardlySet]

And I keep asking, without much response, whether this is something happening to new authors here rather than ones established before the AI problem became a problem. Is the site letting established authors through on this and only putting the thumbs to new author attempts? Does the site, in fact, have enough handle on how to be right most of the time on the call of what is too much AI support (whatever that is) or is it knee-jerk insulting writers on how much originality they are putting into their writing--and only challenging new submitters on it because they realize established writers were doing their thing before AI support was a thing?

It is my impression that mostly new authors are affected by this, although we also had authors who claimed that their years-old stories were pulled for AI usage. I'd say that whoever trips the alarm is being checked for AI even on their past submissions but yeah, I am just speculating based on what some authors posted on the forum.

 

[intim8]

The purpose of this thread is to offer positive suggestions.

In other words, write well. Write emotionally, provacatively, creatively, not mechancially.

 

[KeithD]

This is a good question. I'd like to hear from a) experienced Lit authors if they've encountered an AI rejection, and b) new Lit authors who have NOT encountered this problem.

No, I have not. Over seventeen years and 1,600 stories here I've only been rejected on misperceived underage and the citation of a fake URL when I didn't know that no URLs are accepted--and that's been no more than four or five times. I have not been rejected yet on AI use. If I am, I'll directly e-mail Laurel and give her a chance to let the story through as written (as she did on almost everything else I've had erroneously rejected). If the story still is rejected, I'll stop submitting stories to the site, because, after all this time, both I and Laurel know I'm not using AI support in my writing. Like you, I've gotten good/quick support from Laurel on rejections I've challenged.

 

[intim8]

I'd like to hear from a) experienced Lit authors if they've encountered an AI rejection, and b) new Lit authors who have NOT encountered this problem.

I have seen several in the various threads who have tens of stories posted, then just started getting hit.

 

[KeithD]

I have seen several in the various threads who have tens of stories posted, then just started getting hit.

That's that problem, I think. We're getting rumor and third-party claims. What is needed is "my personal experience here" from new writers/established ones.

 

[nice90sguy]

Don't follow all the rules, and ignore most of the recommendations about your writing, ESPECIALLY @SimonDoom 's O.P, which was excellent, but pretty obviously he made heavy use of Grammarly when he wrote it.

 

[StillStunned]

In other words, write well. Write emotionally, provacatively, creatively, not mechancially.

That's a rather abstract suggestion. I'm sure anyone who's affected by the issue wants tips on *how* to write emotionally, provocatively, creatively.

Also, in response to the other question going around: I've been on Lit for less than a year. I've published 50 stories in that time (of which 20 were removed at my own request), and I've never been rejected for suspected use of AI.

 

[intim8]

What is needed is "my personal experience here" from new writers/established ones.

Yeah, it's third party from me, but in those other threads, it was direct first hand reports. All in the last couple of months. You can go look for them if you want. Those threads are listed below here under "Similar threads"

 

[intim8]

That's a rather abstract suggestion. I'm sure anyone who's affected by the issue wants tips on *how* to write emotionally, provocatively, creatively.

The internet is filled with such tips. There are entire bookstore aisles devoted to books about it.

Point is, it's a good idea regardless of Laurel's choice of AI-detecting AI bot.

 

[SimonDoom]

Don't follow all the rules, and ignore most of the recommendations about your writing, ESPECIALLY @SimonDoom 's O.P, which was excellent, but pretty obviously he made heavy use of Grammarly when he wrote it.

Ignore the other bots. I'm the only bot you can trust.

 

[StillStunned]

Ignore the other bots. I'm the only bot you can trust.

This is a *very* low-budget version of Terminator 2.

 

[djrip]

Because people who learn their writing from the same informative web sources that are used to train AI don't seem to know this.

Every sample I've seen from someone whose story has been rejected seems to follow the same pattern of "piece of information, unassociated piece of information, third random piece of information."

If this thread is about providing tips to avoid sounding like AI, this is where I'd begin. And it might be basic writing to you and me, but we're not the ones getting flagged, because we know this.

But there is no rule stating that sentences must connect together in any particular way. If an authors style is to interject seemingly unrelated thoughts rapid fire, then that is their style. It would have been fine 5 years ago. The only problem with it now is AI detectors falsely declaring it AI.

 

[djrip]

And I keep asking, without much response, whether this is something happening to new authors here rather than ones established before the AI problem became a problem. Is the site letting established authors through on this and only putting the thumbs to new author attempts? Does the site, in fact, have enough handle on how to be right most of the time on the call of what is too much AI support (whatever that is) or is it knee-jerk insulting writers on how much originality they are putting into their writing--and only challenging new submitters on it because they realize established writers were doing their thing before AI support was a thing?

At least two of the people posting about AI rejections have stories posted from years ago. The recent AI conversation matters thread guy, and mypenname3000, iirc. Probably others.

 

[SimonDoom]

This is a *very* low-budget version of Terminator 2.

"Come with me if you want to Lit."

 

[lovecraft68]

I like this thread better than the other one. At least Simon isn't pretending to be the chosen herald for Laurel.

I never thought my hack writing would be an advantage, but apparently meh grammar and characters that speak in slang and butcher the language is a fool proof plan to avoid getting accused of AI

 

[intim8]

If an authors style is to interject seemingly unrelated thoughts rapid fire,

Style isn't about breaking the "rules", it's about bending them, molding them to what you need.

Yes, I know the rules are not fixed in stone, that they evolve, and that breaking them can make a sharp point. But that is selective, and it takes a solid knowledge of them to break them effectively. A lot of new/amateur writers uses "muh style!" to excuse not even knowing the rules.

 

[djrip]

Style isn't about breaking the "rules", it's about bending them, molding them to what you need.

Yes, I know the rules are not fixed in stone, that they evolve, and that breaking them can make a sharp point. But that is selective, and it takes a solid knowledge of them to break them effectively. A lot of new/amateur writers uses "muh style!" to excuse not even knowing the rules.

But there are no rules about this. We're not talking about comma splices, we're talking about how an author chooses to organize their thoughts, or present their points. It is absolutely a valid choice for an author to choose to present a series of statements that don't obviously relate, if they wish. Such a choice would never have been rejected before AI came around and AI detectors started using things like that to identify AI. If they even do; what triggers them is all purely speculation on all of our parts, and varies from detector to detector.

 

[intim8]

we're talking about how an author chooses to organize their thoughts, or present their points. It is absolutely a valid choice for an author to choose to present a series of statements that don't obviously relate

It can be. If that is effective. I'm saying that "my style" does not excuse an ineffective use of such things. And 99.99% of people who claim it as their style are doing just that. The remaining .01% are probably not going to write a whole story in a way that looks like semi-competent AI.

That doesn't mean that AI detector AIs will never hit on writers who do things like that well (assuming the premise here about it being what triggers them), but it supports the OP's proposition that good writing is far less likely to than bad writing.

Unfortunately, that won't last long. But when AI gets to that point, you really have to start to question the need to screen them out in the first place.

 

[Britva415]

I think the problem stems from authors using tools that stifle their own voice and make them sound generic.

I haven't surveyed the specific authors who have prominently protested their rejections on AI grounds, but I see shitloads of stories which have the problem of authors writing generically and not actually having a voice of their own.

Writing robotically in a way which is how they think prose is supposed to sound, without having any sense of whether it serves the story or not. The impression these writings leave me with is that the writer either hasn't read very much diversity of material, or, if they have, haven't thought at all about how styles differ and serve different pieces in different ways.

Language wise, it's basically indistinguishable from AI, which also can't use style to serve the story. Not even when instructed to try for a certain style. The only way to tell it's not AI is that AI can't narrate a coherent sequence of events with a story arc. Humans write this way on Lit so often that deciding whether it's AI or not is not a matter of "voice," unless one isn't actually trying to read the story and is only judging by voice.

*ahem*

... if you see what I mean.

So, there are two sides to this. On the author side, don't write like a robot. On the review side... don't let robots judge the reading? Maybe that will be feasible some day.

 

[Bazzle]

I had a professor at uni who hammered on and on about avoiding "nice" in academic writing. Even now I shy away from it.

It doesn't mean anything. Why? Is the comment regarding "Nice"
Unless it's the city in France. Which is great.

 

[StillStunned]

It doesn't mean anything. Why? Is the comment regarding "Nice"
Unless it's the city in France. Which is great.

It's the most mediocre of compliments. It doesn't have the negative connotations of "fine", at least. On the sliding scale of enjoyment, those two are perhaps at equal distance either side of "neutral".

 

[alohadave]

(for some reason, Grammarly really dislikes the use of the word 'nice')

It hates 'really' as well. Flags every use. Grammarly isn't particularly useful for dialog.

 

[SimonDoom]

If they even do; what triggers them is all purely speculation on all of our parts, and varies from detector to detector.

It's not PURE speculation. I'll grant you that it's some speculation. But there's still some useful information we can work with. Some authors do not have this problem. Some, who do, from what I can see manifest certain writing patterns that, to me, read like AI. So I think the point stands that there are things an author can do to reduce the chance of being flagged for AI. It may seem wrong or unfair, and I won't argue with that, but it is what it is, and one can complain, or one can adapt and try to find a workaround.

 

[intim8]

Some authors do not have this problem. Some, who do, from what I can see manifest certain writing patterns that, to me, read like AI.

You know the story of the bombers that came back full of bullet holes?

We're only seeing the strories that got through the screening.